Individual Differences
Individual Differences
I. OVERVIEW
The study of individual and group differences in psychological traits is the domain of differential psychology. The measurement of such differences has led to the accumulation of a vast array of descriptive data that is of direct scientific and practical interest. A more fundamental aim of differential psychology, however, is to provide one approach to an understanding of behavior. Differential psychology approaches this task through a comparative analysis of behavior under varying environmental and biological conditions. By relating observed behavioral differences to known concomitant circumstances, we are able to identify the relative contributions of different factors to behavior development. If we can discover why one individual reacts differently from another, we shall thereby advance our knowledge of what makes each behave as he does.
Individual differences in behavior are not limited to the human species; they occur throughout the animal scale. Investigations of animal behavior, from unicellular organisms to anthropoid apes, reveal wide individual differences in learning, motivation, emotionality, and other measurable traits. So large are these differences that the distributions of individual performance overlap, even when widely separated species are compared. When tested with the same learning problem, for example, the brightest rat in a given sample may excel the dullest monkey (see Anastasi 1958, pp. 48-53).
Interaction of heredity and environment
All traits are the result of innumerable and complex interactions between the individual’s heredity and his environment. An individual’s heredity consists of the genes he receives from each parent at conception. Genes are units of complex chemical substances that affect the course of the individual’s development from the one-cell stage to the mature organism. If there is a chemical deficiency or imbalance in one of these genes, a seriously defective organism may result, with bodily anomalies as well as severely retarded intelligence. Certain rare forms of mental deficiency, such as that associated with phenylketonuria, have been traced to defective genes. In these cases, some of the minimum physical prerequisites for normal intellectual growth are lacking. Except for such pathological deviates, however, heredity sets very broad limits to behavior development. Within these limits, what the individual actually becomes depends upon his environment.
Environment includes the sum total of stimuli to which the individual responds from conception to death. It comprises a vast multiplicity of factors, ranging from air and food to the social and emotional climate of home and community and the beliefs and attitudes of one’s associates. Environmental influences begin to operate before birth. Nutritional deficiencies, toxins, and other chemical or physical conditions of the prenatal environment may exert a profound and lasting effect upon both physical and mental development. Several varieties of mental deficiency, for example, result from abnormalities of prenatal environment. Such terms as “inborn,” “innate,” and “congenital” may be misleading because they suggest that all characteristics present at birth are hereditary, which is not the case. Another common confusion is that between organic and hereditary conditions. Mental deficiency resulting from early brain injury, for instance, may be properly said to have an organic but not a hereditary origin.
The research techniques employed to investigate the operation of hereditary and environmental factors in behavioral development may be classified under three major approaches, namely, selective breeding, experimental variation, and statistical studies of familial resemblance. [SeeGenetics, article onGenetics and Behavior.]
Selective breeding for behavioral characteristics has been successfully applied to several species. From a single initial group of rats, for example, it proved possible to breed two strains comprising good and poor maze-learners, respectively (Tryon 1940). After seven generations during which good performers were mated only with good performers and poor performers only with poor performers, virtually no overlapping remained between the distributions of maze scores of the two strains. Later cross-breeding of good and poor performers yielded results indicating that a large number of genes influence maze performance in rats. That the two strains did not differ in general learning capacity, however, was demonstrated by the finding that both strains performed equally well in certain other learning problems (Searle 1949). Other data suggested that emotional and motivational factors may have played an important part in the maze performance of the two strains. There is also evidence that the two strains differed significantly in certain biochemical factors that affected the efficiency of brain action (Rosenzweig et al. 1960).
More recent selective breeding experiments have extended these procedures to other behavior functions and other types of organisms (see Fuller & Thompson 1960). Of particular significance was the development of techniques for measuring individual differences in behavior among such organisms as the fruit fly Drosophila (Hirsch 1962). It thus became possible to capitalize on the mass of available genetic knowledge regarding the morphology of Drosophila, as well as on such other advantages as the short time span between generations and the abundance of progeny. By these procedures, a strain of fruit flies has been bred that will fly toward a source of light and another that will fly away from it. Similarly, one strain has been developed that tends to fly upward, and another downward, when released in a vertical maze. [SeeEugenics.]
A second approach to the heredity-environment problem is concerned with the behavioral effects of systematic variations in experience. This method has frequently been used with animals to study a wide variety of activities, ranging from the swimming of tadpoles and the singing of birds to sexual behavior and care of the young (see, for example, Beach & Jaynes 1954). Many experiments have utilized artificial devices to reduce or cut off sensory stimulation or to eliminate opportunity for the exercise of specific motor functions, in order to discover how far a function will develop in the absence of appropriate experience. Others have followed the opposite approach, providing intensive controlled training in various activities. Significant effects of such prior experiences have been reported for nearly all aspects of behavior, including perceptual, motor, learning, emotional, and social reactions. Through such experiments, many activities formerly regarded as completely unlearned or “instinctive,” such as nest building and care of the young by rats, have been shown to depend upon the animal’s prior experiences (Beach & Jaynes 1954; Birch 1956). Even when the animal has no opportunity to learn the specific activity in question, his behavior may be influenced by the exercise of other, related functions. [Seeethologyandinstinct.]
Of particular interest to differential psychology is a group of experiments on monkeys demonstrating the effect of prior experience upon learning ability itself (Harlow 1949). Through the formation of learning sets, the animals were able to learn the solution of complex problems because of their prior experience in solving simpler problems of a similar nature. By means of this problem-solving experience, the animal thus “learns how to learn.” Other research has shown that animals exposed to a rich variety of perceptual experience during early life are better subsequent learners than those deprived of such experience.
In studies of infants and young children, one group of experiments has utilized the method of co-twin control. In these experiments, one member of a pair of identical twins is given intensive training in some activity, such as climbing stairs or manipulating objects, while the other is retained as a control subject and temporarily prevented from exercising the function under investigation. The results generally show that, if training is introduced when the child is physically ready for it, progress will be faster than if training is given earlier. Another method is the comparative study of children reared in culturally deprived or psychologically limited environments, such as isolated mountain communities, gypsy caravans, houseboats, and orphanages. Considerable intellectual retardation has been found in all these situations, the retardation becoming more severe with increasing age. Some studies have demonstrated significant improvement in intellectual functioning as a result of preschool education on the part of orphanage children and other culturally deprived groups.
Among adults, it is well established that intelligence test scores correlate highly with the amount of schooling completed. Follow-up studies of groups retested after 10 to 30 years have revealed that individuals who continue their education longer show larger mean gains in intelligence test scores than do those with less intervening education (see Anastasi 1958, chapters 4 and 7).
The third major approach is based upon statistical analyses of familial resemblances and differences. Similarity of performance on both aptitude and personality tests has been investigated for parents and children, siblings, and twins (including both fraternal and identical pairs). In general, the closer the hereditary relation, the more similar the test scores will be. On most intelligence tests, for example, identical twin correlations are close to .90, being nearly as high as the correlations between test and retest scores of the same persons. Fraternal twin correlations cluster around .70; those between siblings cluster around .50, as do those between parents and children. It should be noted, however, that a family is a cultural as well as a biological unit and that a hierarchy of environmental similarity and mutual influence parallels the hierarchy of hereditary relationships. Special studies of foster children and of identical twins reared apart permit some isolation of hereditary and environmental influences, but various uncontrolled factors preclude definitive interpretation of results (see Anastasi 1958, chapter 9).
Physique and behavior
The term “physique” is herein used to refer collectively to all organic characteristics, including anatomical, physiological, and biochemical properties of the organism. The relationship between physique and behavior concerns the differential psychologist for both theoretical and practical reasons. Investigation of the role of physical factors in the development of psychological traits advances our understanding of the causes of individual differences in these traits. Insofar as heredity contributes to behavioral differences, moreover, identification of the physical bases of such differences is an important step in tracing the long and intricate path from gene to behavior. From a practical standpoint, interest in the relationship between physical and psychological traits stems from the possibility of assessing people and predicting behavior on the basis of physical characteristics.
In evaluating the results of any study of the relationship between physique and behavior, we must bear in mind that a significant correlation may mean that the physical condition influences the behavior in question, or that the behavior influences the physical condition, or that both result independently from the common influence of some third factor. [SeePsychology, article onconstitutional psychology.] The many ways in which physical factors influence behavior may be arranged along a continuum, ranging from relatively direct and rigidly limiting control to highly indirect and flexible relationships (see Anastasi 1958, chapter 5). The former extreme is illustrated by neurological, glandular, and metabolic disorders (of either hereditary or environmental origin) that lead to severe behavioral pathology. The abnormally small brain of the microcephalic, the underactive thyroid of the cretin, and the metabolic disorder of the phenylketonuric are examples of this mechanism. In all these instances, the individual lacks the minimum organic prerequisites for normal intellectual development. Unless the physical disorder can be corrected at an early developmental stage, behavioral deficiencies will result.
A more indirect influence of physique on behavior is illustrated by severe sensory or motor handicaps, such as blindness, deafness, or cerebral palsy, which reduce normal social intercourse and interfere with education. Unless special instructional techniques are employed, these physical handicaps may seriously retard intellectual development and affect personality in various ways. Depending upon concomitant circumstances, however, these physical conditions may lead to very dissimilar effects in different individuals.
At the other extreme of the continuum of indirectness is the operation of social stereotypes. Through this mechanism, the individual’s visible physical characteristics serve as social stimuli, which elicit differential responses from his associates. They may thus affect the attitudes he encounters, the opportunities he receives, and the shaping of his own self concept. As a result, his behavior may gradually come to approximate that associated with the stereotype.
The reverse relation, in which behavior influences physique, is illustrated by the powerful shoulder muscles of the swimmer and by the scholar’s stoop. Smiles and frowns, too, eventually leave their marks upon the human countenance. Of particular interest in this connection are psychosomatic disorders, that is, physical diseases in whose development psychological factors play at least a contributing part. From another angle, research on immigrant groups has demonstrated that such characteristics as stature and skull shape may be influenced by dietary habits, child-rearing practices, and other culturally determined behavior. Still another example is provided by comparative studies of schizophrenics and normals, in which organic differences may result from differences in emotional stress, degree of activity, nutritional state, and other behavioral variables associated either with the psychotic condition itself or with institutionalization.
The third type of causal relation between physique and behavior is that in which the correlation results from the common influence of a third factor, such as socioeconomic level. The child reared in a superior home, for example, has richer opportunities for intellectual development, as well as better diet, hygiene, and medical care, than does the child who grows up in a city slum or poor rural area. Consequently, some positive correlation will be found between intelligence and a number of physical conditions within a culturally heterogeneous population. The correlation usually disappears, however, when socioeconomic level is held constant (see Anastasi 1958, chapter 5).
Nature and distribution of intelligence
Intelligence has been commonly identified with the intelligence quotient (IQ) obtained on an intelligence test. Such tests do reflect, at least partly, the concept of intelligence current in the culture in which they were developed. Nevertheless, once an intelligence test has undergone the years of preparation and standardization required for its construction, it may tend to freeze a particular concept of intelligence and thereby retard change. Moreover, intelligence tests are designed to meet practical demands within specific settings. Hence they often represent a compromise between practical testing needs and the concept of intelligence that might have developed in the less restricted context of basic research. [SeeIntelligence andintelligence testing.]
Most intelligence tests measure chiefly scholastic aptitude, or that combination of abilities required for school achievement. Modern intelligence testing originated with Alfred Binet’s development of a test to assess intellectual retardation among school children. Current intelligence tests have frequently been validated against such academic criteria as school grades, teachers’ ratings of intelligence, promotion and graduation data, and amount of schooling completed. High positive correlations between test scores and these criterion measures are regarded as evidence that the test is a valid measure of intelligence.
With regard to content, most intelligence tests are heavily weighted with verbal aptitudes. To a lesser extent, they may also sample arithmetic skills, quantitative reasoning, and memory. Different intelligence tests, moreover, may cover somewhat different combinations of abilities. Nonlan-guage and performance tests, for instance, often make much heavier demands upon spatial visualization, perceptual speed and accuracy, and abstract reasoning than do the usual verbal-type tests. It is largely for this reason that an IQ should always be accompanied by the name of the test from which it was derived. [SeeAchievement testing.]
Following the widespread application of intelligence tests, psychologists soon recognized that certain aptitudes remained largely unmeasured by these tests. The increasing participation of psychologists in vocational counseling and in the screening and classification of industrial and military personnel highlighted the need for tests of other aptitudes. As a result, so-called special aptitude tests were developed to supplement general intelligence tests in mechanical, clerical, musical, artistic, and other aptitude areas. [SeeAptitude Testingandvocational interest testing.]
At the same time, clinical psychologists working intensively with individual cases were impressed with the large intraindividual differences often found from one intelligence test to another, or among different parts of the same intelligence test. Thus an individual might score consistently better on performance than on verbal tests; or within the same intelligence test, he might do well on numerical and poorly on verbal tasks.
Concurrently, basic research on the nature of intelligence was being conducted by the techniques of factor analysis. Essentially, these techniques involve statistical analysis of the intercorrelations among test scores in the effort to discover the smallest number of independent factors that can account for their interrelations. Among the aptitudes or “factors” thus identified are verbal comprehension, word fluency, arithmetic skills, quantitative reasoning, perceptual speed, spatial visualization, and mechanical comprehension. Through factor analysis, what had formerly been called intelligence could itself be subdivided into relatively independent aptitudes, and these aptitudes could be recombined with some of those underlying special aptitude tests to provide a more comprehensive picture of intelligence.
Later factor-analytic research has been extending the concept of intelligence in our culture to include creativity, originality, and divergent thinking, in contrast to the comprehension and retention skills emphasized in traditional intelligence tests (Guilford 1959). Several research projects concerned with creativity have been exploring new testing areas and developing new types of tests suitable for measuring divergent thinking in both children and adults. Although these new test materials have not been incorporated to any appreciable extent in commercially available intelligence tests, it is likely that they will be utilized increasingly in the intelligence tests of the future. [SeeFactor analysis.]
Recognizing that the IQ is a composite measure and that its nature shifts with changing concepts of intelligence, we may nevertheless inquire into its distribution in the population. In intelligence, as in all psychological traits, individuals do not fall into sharply separated categories or types. Instead they vary by degree along a continuous scale. Most psychological tests are constructed so as to conform with the bell-shaped normal probability curve, with the greatest clustering of persons near the center of the range and a gradual decrease in numbers as the extremes are approached. First derived by mathematicians in their study of probability, the normal curve is obtained whenever the variable measured is the result of a very large number of independent and equally weighted factors. In view of the extremely large number of genes and of environmental factors that contribute to the development of intelligence in the general population, the normal curve is generally accepted as the most appropriate model for the distribution of intelligence.
The mentally deficient and the gifted represent the lower and upper extremes of the distribution of intelligence. Because the distribution is continuous, there is no clearly deined separation between these groups and the normal. In terms of intelligence test performance, mental deficiency is customarily identified with IQ’s below 70. About 2 to 3 per cent of the general population fall in this range. Decisions regarding the disposition and treatment of individual cases ought, of course, to take much more than an IQ in account. They should be based upon a comprehensive study of the individual’s intellectual level, social competence, educational history, physical condition, familial situation, and other pertinent data. [SeeMental Retardation.]
With regard to etiology, one distinction is that between unifactor and multifactor mental deficiency. Unifactor cases are extreme deviates; they manifest both organic pathology and intellectual defect, which are traceable to a single defective gene or to a major environmental disturbance. Multifactor defectives, on the other hand, represent merely the lower end of the normal distribution of intelligence. They exhibit varying degrees of deficiency, depending upon the particular combination of adverse hereditary and environmental factors in each case. Since the unifactor defectives are added to the multifactor defectives at the lower end of the distribution, the frequency of low IQ’s should exceed that expected in a mathematically derived normal probability curve. Large-scale surveys of the distribution of IQ’s in various populations do, in fact, reveal such a deviation from normality at the low end of the scale (Dingman & Tarjan 1960; Roberts 1952).
The intellectually gifted have been investigated by many techniques and from many points of view. In the monumental study conducted by Terman and his associates at Stanford University, approximately one thousand California school children with Stanford-Binet IQ’s of 140 or higher were intensively examined and followed up through adulthood (Terman & Oden 1959). Slightly more than 1 per cent of the general population obtain IQ’s at this level. The results of the Stanford study, which have been corroborated in other studies conducted elsewhere, revealed the gifted child as typically successful in school, healthy, emotionally well-adjusted, having wide interests, and excelling his average classmates in nearly every trait measured. Although there were individual exceptions, the group as a whole clearly dispelled the early stereotype of the intellectually gifted child as weak, sickly, timid, and narrowly specialized. As they grew into maturity, the California gifted children amply fulfilled their youthful promise in outstanding adult achievements. [SeeTerman; creativity, article ongenius and ability.]
Since the middle of the twentieth century, the most conspicuous development in the investigation of superior deviates is to be found in research on creativity with both adults and children (Golann 1963). Although such research is yielding a wealth of data with important practical and theoretical implications, we must guard against exaggerating the distinction between creativity and intelligence as traditionally measured. To be sure, neither a high IQ on current intelligence tests nor high academic achievement is identical with creativity. These qualifications do not ensure that an individual will make outstanding contributions. On the other hand, they do not preclude creativity, nor are they completely unrelated to it. Traditional intelligence tests show a moderate but significant positive correlation with measures of creativity. Rather than differentiating between intelligence and creativity, moreover, we need to broaden the concept of intelligence to include newly identified creative traits (Guilford 1959). [SeeCreativity, article onpsychological aspects.]
Sex differences
An important fact about all comparisons among human groups is that individual differences within each group are far greater than average differences between groups. As a result, the distributions of the groups overlap to a marked degree. Even when the means of two groups differ by a large and statistically significant amount, individuals can be found in the low-scoring group who surpass individuals in the high-scoring group. Hence an individual’s membership in a particular group is a very unreliable indicator of his standing in most psychological traits. Group averages need to be evaluated with reference to some measure of overlap of total distributions, such as the proportion of one group that reaches or exceeds the median of the other.
Another methodological problem in group comparisons arises from the use of unrepresentative samples, in which selective factors may have operated differentially for the two populations. Insofar as more boys than girls drop out of school, for example, a comparison of the intelligence test performance of high school boys and girls will yield a mean sex difference in favor of boys. This difference would disappear, however, if we were to include drop outs, who tend to score near the low end of the distribution. A similar error in the opposite direction is illustrated by surveys of institutions for mental defectives, which generally show an excess of males. Although once regarded as evidence of the greater incidence of mental deficiency among males, these findings were later traced to selective admission policies. For a variety of social and economic reasons, mentally defective women are more likely to remain in the community than are males of the same intellectual levels.
The use of global scores on intelligence tests may also be misleading in the study of sex differences. In the construction of several intelligence tests, such as the Stanford-Binet, sex differences have been ruled out by omitting items that favored either sex. Even when this is not done, a composite score on a heterogeneous test may obscure genuine group differences in specific abilities.
Psychological test surveys in contemporary Western cultures have demonstrated significant mean differences between the sexes in a number of aptitudes and personality traits (see Anastasi 1958, chapter 14). Males as a group excel in speed and coordination of gross bodily movements, spatial orientation, mechanical comprehension, and arithmetic reasoning. Females excel in manual dexterity, perceptual speed and accuracy, memory, numerical computation, verbal fluency, and other tasks involving the mechanics of language. Among the principal personality differences found between the sexes are the greater aggressiveness, achievement drive, and emotional stability of the male and the stronger social orientation of the female.
Most investigations of sex differences yield only descriptive data about existing differences between men and women within a given culture. The origins of such differences must be sought in the complex interactions of cultural and biological factors. Although living in the same homes, boys and girls in most societies are reared in different subcultures. In countless ways, they receive differential treatment from parents, other adults, and age peers. They are dressed differently, given different toys, taught different games, and expected to behave differently in many situations. The personalities of mother and father are themselves important factors in the child’s developing concept of sex roles, providing models of what is expected of each sex in the particular culture.
From a biological viewpoint, the different parts men and women play in the reproductive function undoubtedly contribute to sex differentiation in psychological development. Thus the long period of child bearing and child rearing, which falls biologically upon the female, has far-reaching implications for sex differences in interests, attitudes, emotional traits, vocational goals, and achievement. Sex differences in aggressiveness and dominance are associated with the greater body size, strength, and physical endurance of the male, as well as with the presence of male sex hormones.
Another significant sex difference is to be found in the developmental acceleration of girls. Not only do girls reach puberty earlier than boys, but throughout childhood they are further advanced toward their own adult status in all physical traits. The psychological effects of this sex difference in developmental rate probably vary widely from trait to trait. In infancy, the developmental acceleration of girls may be an important factor in their more rapid acquisition of language and may give them a head start in verbal development as a whole. These few examples suffice to illustrate the varied and intricate mechanisms whereby biological and cultural differences between the sexes interact to produce differences in aptitudes, interests, and other psychological traits.
Race and culture
The biological concept of race refers to populations that differ in the relative frequency of certain genes. Races are formed when a group becomes relatively isolated, for either geographic or social reasons, so that marriage among its members is more common than marriage with outsiders (see Dobzhansky 1962, chapter 10). The very isolation that leads biologically to race formation also fosters cultural differentiation. Hence the populations that have been compared in research on race differences have usually differed in their cultural environments as well. Under these conditions, any differences in aptitudes or personality traits found between racial groups cannot be unequivocally attributed to racial or genetic factors.
In an effort to circumvent cultural differences among groups, some investigators have utilized so-called culture-free tests (see Anastasi 1958, pp. 561-569; 1961, chapter 10). These tests have been especially developed for comparative studies of persons reared in different cultures. They not only undertake to eliminate language barriers, but they also employ content presumably common to all cultures. Parenthetically, it should be noted that the term “culture-free test” is a misnomer. Since a psychological test is only a standardized measure of a behavior sample, any condition that influences behavior will be reflected in test scores. It is theoretically possible to construct a test that presupposes only experiences common to different cultures, but not one that is free from cultural influences.
Even this theoretical goal, however, has not been attained by any available test; each test still favors certain cultural groups and handicaps others. Every psychological test inevitably draws most heavily upon the information, skills, work habits, and attitudes fostered by the culture in which it was developed. The difference between “culture-free” and other tests is merely one of degree. Even the use of paper and pencil or the presentation of abstract tasks having no immediate practical significance will tend to discriminate against some cultures or subcultures. Other cultural differences include intrinsic interest of test content, degree of familiarity with pictorial or diagrammatic representation, rapport with the examiner (especially one of a different race), motivation to perform well on a test, competitive desire to excel others, and previously developed problem-solving attitudes.
The analysis of population differences has also been approached through experimental designs that permit some isolation of biological and cultural influences. Test performance of hybrid, or racially mixed, groups, has been investigated for this purpose. It has been argued that, if one race is intellectually superior to another because of genetic factors, the hybrid offspring of both races should be intermediate in intelligence. Genetically, this hypothesis is questionable, since it assumes complete linkage between the genes determining skin color or other racial indices and the genes determining intelligence. With incomplete linkage, the correlation between racial characteristics and intelligence would disappear within a few generations of crossbreeding. The results are further complicated by the fact that race mixture is usually selective within either or both races, as well as by the tendency toward greater cultural assimilation of hybrids. In groups that are fairly homogeneous in their assimilation of the dominant culture, the correlation between test score and extent of race mixture is negligible (e.g., Rohrer 1942).
Another group of studies concerns changes in the comparative test performance of racial groups with age. Several investigations of American Negro infants and preschool children, for example, revealed little or no retardation in terms of white norms (Anastasi & D’ Angelo 1952; Brown 1944; Gilliland 1951; Pasamanick 1946). Tests of school-age children conducted in the same areas and periods, on the other hand, showed significant mean retardation that increased with age. These findings are similar to those obtained with other culturally deprived groups. The age decrement has been ascribed to the cumulative effects of deficient environments and to the increasing inadequacy of such environments to meet the expanding intellectual needs of the growing child. From a broader viewpoint, such an age decrement in relation to test norms may be said to occur when a test demands intellectual functions not fostered in a particular culture or subculture (Levinson 1961).
A third approach is based upon a comparison of samples of the same race reared in different environments. In general, such studies have yielded larger differences in test performance among subgroups of a given race living in different milieus than among different racial groups living under more nearly similar conditions (see Anastasi 1958, pp. 584-592). That the regional differences found within a racial population are associated with cultural differences rather than with selective migration has been demonstrated in several studies. Of particular relevance are the results of a longitudinal investigation of American Negro children who had moved from an area with poorer school facilities to one with better school facilities (Lee 1951). Mean intelligence test scores of these children improved significantly with increasing length of residence in the educationally more favorable area.
Despite a mass of descriptive data on psychological differences among races, research on the origins of such differences is meager and beset with methodological difficulties. In the light of available knowledge, only a few conclusions can be drawn with confidence. First, no biological basis has as yet been identified for any existing psychological differences among races. Second, there is considerable evidence, both from racial studies and from other types of investigations in differential psychology, showing the part played by cultural factors in producing the sort of behavioral differences commonly found among racial groups. Finally, in all psychological traits, the range of individual differences within each race is far greater than mean differences between races.
Anne Anastasi
[Other relevant material may be found indevelopmental PSYCHOLOGY; INTELLECTUAL DEVELOPMENT; personality, article onpersonality development; traits.]
BIBLIOGRAPHY
Anastasi, Anne 1958 Differential Psychology: Individual and Group Differences in Behavior. 3d ed. New York: Macmillan. ? First published in 1937.
Anastasi, Anne 1961 Psychological Testing. 2d ed. New York: Macmillan. ? First published in 1954.
Anastasi, Anne; and D’Angelo, Rita Y. 1952 A Comparison of Negro and White Preschool Children in Language Development and Goodenough Draw-a-Man Iq. Journal of Genetic Psychology 81: 147-165.
Beach, Frank A.; and Jaynes, Julian 1954 Effects of Early Experience Upon the Behavior of Animals. Psychological Bulletin 51:239-263.
Birch, Herbert G. 1956 Sources of Order in the Maternal Behavior of Animals. American Journal of Orthopsychiatry 26: 279-284.
Brown, Fred 1944 An Experimental and Critical Study of the Intelligence of Negro and White Kindergarten Children. Journal of Genetic Psychology 65: 161-175.
Dingman, Harvey F.; and Tarjan, George 1960 Mental Retardation and the Normal Distribution Curve. American Journal of Mental Deficiency 64: 991-994.
Dobzhansky, Theodosius 1962 Mankind Evolving: The Evolution of the Human Species. New Haven: Yale Univ. Press.
Fuller, John L.; and Thompson, W. Robert 1960 Behavior Genetics. New York: Wiley.
Gilliland, Adam R. 1951 Socioeconomic Status and Race as Factors in Infant Intelligence Test Scores. Child Development 22:271-273.
Golann, Stuart E. 1963 Psychological Study of Creativity. Psychological Bulletin 60:548-565.
Guilford, J. P. 1959 Three Faces of Intellect. American Psychologist 14:469-479.
Harlow, Harry F. 1949 The Formation of Learning Sets. Psychological Review 56:51-65.
Hirsch, Jerry 1962 Individual Differences in Behavior and Their Genetic Bases. Pages 3-23 in Eugene L. Bliss (editor), Roots of Behavior: Genetics, Instinct, and Socialization in Animal Behavior. New York: Harper.
Lee, Everett S. 1951 Negro Intelligence and Selective Migration: A Philadelphia Test of the Klineberg Hypothesis. American Sociological Review 16:227-233.
Levinson, Boris M. 1961 Subcultural Values and IQ Stability. Journal of Genetic Psychology 98:69-82.
Pasamanick, Benjamin 1946 A Comparative Study of the Behavioral Development of Negro Infants. Journal of Genetic Psychology 69:3-44.
Roberts, John A. Fraser 1952 The Genetics of Mental Deficiency. Eugenics Review 44:71-83.
Rohrer, John H. 1942 The Test Intelligence of Osage Indians. Journal of Social Psychology 16:99-105.
Rosenzweig, Mark R.; Krech, David; and Bennett, Edward L. 1960 A Search for Relations Between Brain Chemistry and Behavior. Psychological Bulletin 57:476-492.
Searle, Lloyd V. 1949 The Organization of Hereditary Maze-brightness and Maze-dullness. Genetic Psychology Monographs 39:279-325.
Terman, Lewis M.; and Oden, Melita 1959 Genetic Studies of Genius. Volume 5: The Gifted Group at Mid-life: Thirty-five Years’ Follow-up of the Superior Child. Stanford Univ. Press.
Tryon, Robert C. 1940 Genetic Differences in Maze-learning Ability in Rats. Volume 1, part 1, pages 111-119 in National Society for the Study of Education, 39th Yearbook. Bloomington, III.: Public School Publishing Company.
II. SEX DIFFERENCES
Since the beginning of the twentieth century, psychological sex differences have been the subject of continuous research. The objectives and orientation of this research, however, have shifted several times. At the beginning, the feminist movement generated interest in the question of whether the intelligence of women was or was not equal to that of men. What most investigators hoped to find was scientific evidence for the equality of the sexes.
As techniques for measuring aspects of personality as well as mental abilities were developed in the 1920s and 1930s, more and more comparisons of male and female groups on nonintellectual characteristics were reported. Psychoanalysis was becoming increasingly influential during this period, and personality theorists drew from the writings of Freud and his followers hypotheses about what sex differences should be explored through research. The social objectives were to prevent neurosis and improve relationships between the sexes through understanding the differing emotional needs and ways of expression in men and women. Some investigators also hoped during this period—roughly the second quarter of the century—to construct a scale for measuring general masculinity or femininity that would be accurate enough to differentiate between persons of the same sex. They thought that such a scale would have many practical uses, such as the assignment of individuals to suitable occupations and the diagnosis of homosexual trends.
During the 1950s the emphasis shifted once more. Investigators became very much aware of sex roles. Questions about when and how a young child learns what these roles are and develops preferences and patterns of behavior in accordance with them began to seem very important. The concept of identification, as interpreted in psychoanalytic writings and in other types of personality theory, became the focus of much research effort. The objectives of the research of this period were not just to understand sex differences for their own sake, but to utilize information about this highly visible aspect of personality development as a source of clues about the developmental processes through which many other aspects of personality may have come into existence.
One persistent question has stimulated research interest throughout all these periods: why have women’s achievements failed to match those of men? Why are there so few outstanding female artists, scientists, or statesmen? As the processes of rapid social change have produced an occupational situation in which highly trained professional persons are in demand and unskilled workers are needed in fewer and fewer numbers, this problem has taken on a new urgency. Findings from all the main types of research on differences in abilities, personality traits, roles, and development have been brought to bear on this issue.
Sex differences in abilities
Intelligence
As sophistication in mental-testing procedures increased over the years, it became apparent that the question investigators first asked about the differences in intelligence between the sexes is unanswerable, at least by present methods of measurement. Hundreds of studies were made, most of them using students as subjects. Some reported male superiority, some female superiority, and some no significant differences. Analyzing these conflicting results, one could see that some of the discrepancies arose from selective factors in the groups tested. But, more important, discrepancies were related to the type of intelligence test the investigators happened to use. Some tests consistently gave girls a slight advantage; others favored boys.
From many sources, evidence was accumulating that intelligence tests of all varieties do not and cannot measure pure native capacity. Although genetic differences in intellectual potential undoubtedly exist, the only way we have to test a person is to ask him questions or give him problems to solve. His performance in such a situation always reflects the experience he has undergone as well as his native capacity. It is, therefore, only natural that males show superiority in dealing with some kinds of questions and tasks and that females show superiority in dealing with others. In order to produce intelligence tests that will be equally fair to members of both sexes, psychologists try to include approximately the same number of both kinds of items. McNemar (1942) has explained how this was done in developing the Terman-Merrill revision of the Stanford-Binet test. When this kind of test is given to samples of the population that are really representative, sex differences in over-all score or IQ turn out to be negligible. The two major surveys undertaken by the Scottish Council for Research in Education in 1939 and 1949, in which all children in the country who had been born on certain dates were tested, revealed almost identical mean IQ’s the first time and a 4-point difference in favor of boys the second time. This 4-point difference, although statistically significant for this large number of cases, appears not to have much practical significance in view of the fact that the group test given in the same survey produced a 2-point difference in favor of girls. [SeeAchievement TESTING; Intelligence
and intelligence testing.]
Verbal, mathematical, and spatial abilities
More meaningful than the question of whether one sex is more intelligent than the other is the question of what special abilities are related to sex. Much of the research on this problem is summarized by Terman and Tyler (1954). Females appear to excel in verbal ability, when this is defined as fluency—the use of words rather than the comprehension of verbal meanings. Females tend to talk more, to read faster, and to be less susceptible to reading and speech difficulties. They do not typically score higher than boys on vocabulary tests, and one study of a large and representative sample of English children from 5 to 15 years of age showed consistent although small differences in favor of boys on four oral vocabulary tests from leading intelligence scales (Dunsdon & Roberts 1957).
Males consistently score higher than females on mathematical ability, when this is defined as reasoning or problem solving rather than computational skill. There also is a large and significant difference in favor of males on tests of spatial visualization. These differences, especially the one in mathematics, partially explain why males tend to do better in science. In the annual science talent search carried on each year in the United States, high school male applicants score significantly higher on the achievement test than female applicants, despite the fact that the girls who apply are much more highly selected than the boys who apply (Edgerton & Britt 1947).
Vocational aptitudes
Vocational psychologists have found that on tests of aptitude for particular kinds of work, females score significantly higher for clerical aptitude and dexterity, males significantly higher for mechanical aptitude. Aptitude tests for art and music tend to give higher scores for females, but this may be because more girls than boys are exposed to art and music lessons as they grow up.
Two general findings must be kept in mind if misinterpretations of these differences in ability are to be avoided. One is that in all such comparisons there is a large amount of overlap between the distributions for the two sexes. In making any practical decision about job placement, admission to professional schools and training programs, and the like, one must judge on the basis of the individual’s own ability rather than on the basis of sex. Individual differences far outweigh sex differences. The other general finding from comparisons of sex groups at different ages is that most of these varieties of differences in ability do not show up until the elementary school years or later. With regard to mathematical ability, for example, girls at the kindergarten and preschool levels do as well as boys in counting and identifying numbers. Sex differences become more apparent with increasing education, even when it is coeducation. [SeeAptitude testing.]
Cognitive style
One promising line of research has identified what may be important sex differences in cognitive style. Witkin and others (1954) discovered that females are less able than males to disregard the visual field in which a perceptual pattern they are trying to grasp is embedded. The difference was apparent on tests in which the subject was required to straighten a tilted chair in a tilted room and thus to separate kinesthetic cues from visual distractions; the difference was also apparent on the purely visual embedded-figures test. This tendency toward what Witkin labeled field-dependence was related to personality characteristics having to do with passive acceptance rather than active coping with one’s environment. Sandström (1953) reported a somewhat similar phenomenon. Women are less accurate than men in pointing to a luminous spot of light in a completely darkened room and are more likely than men to show disoriented behavior. In problem-solving experiments, the difficulty females have been shown to experience in restructuring the problem situation would appear to involve this same factor of field-dependence.
Interests, motives, and personality
All of the evidence seems to point to a conclusion that the sexes differ far more in their general orientation to life than they do in abilities.
Interests and values
Strong (1943) has presented the most comprehensive findings with regard to interest differences, findings that are corroborated by other research on the topic. Males respond with a higher degree of preference than do females to interest items of the following kinds: (1)mechanical and scientific activities; (2) physically strenuous, adventuresome activities; (3) legal, political, and military occupations; (4) selling; (5) certain forms of entertainment, such as smokers, “roughhouse” initiations, and chess; (6) certain miscellaneous aspects of work, such as outdoor activity rather than indoor and self-employment rather than working for others. The distinctly feminine interests are seen in responses that indicate greater preference for the following kinds of items: (1) musical, artistic activities; (2)literary activities; (3) certain kinds of persons, especially the unfortunate and disagreeable; (4) certain forms of entertainment, such as for-tunetelling, full-dress affairs, and social-problem movies; (5) clerical work; (6) teaching; (7) social work; (8) merchandise, that is, looking at shop windows, displaying merchandise, etc. Strong constructed a masculinity-femininity scale for the Vocational Interest Blank made up of these discriminating items.
Although representative groups of men and women reveal different patterns of scores on the occupational scales of the Strong and other interest tests, there is some evidence that highly selected groups of women in predominantly male occupations, such as medicine and life insurance selling, obtain interest scores very similar to those of their male colleagues. Women, in general, show a standard pattern of occupational interest scores. The large majority of them receive their only high scores on the office worker, stenographer-secretary, and housewife scales of the Strong Vocational Interest Blank for women. [SeeVocational Interest testing.]
When groups of men and women are compared on the Allport-Vernon (now Allport-Vernon-Lindzey) test of values, men obtain significantly higher scores on theoretical, economic, and political values; this indicates that they are more oriented than women toward abstract ideas, practical success, and power. Women receive higher scores on aesthetic, social, and religious values; this indicates that they are oriented more toward art, religion, and social welfare.
Studies of children. Numerous studies of children’s interests using many observational and assessment techniques reveal marked interest differences even at any early age. Boys engage more in active games and vigorous physical activity and prefer tales of adventure and violent action in books and in radio and television programs. Girls are more likely to enjoy dolls, paper activities, and games calling for skillful movements and to prefer sentimental and domestic stories. Tyler’s comparison of the responses of English and American children to questions on an interest inventory (1956) and Gaier and Collier’s comparison of the reading interests of American and Finnish children (1960) both suggest that sex differences are greater than nationality differences, at least within Western culture.
Motivation
Another large-scale research program that has repeatedly revealed sex differences is the work of McClelland, Atkinson, and their associates on achievement motivation (McClelland et al. 1953). The projective method employed in these studies, the evaluation of stories written about achievement-oriented pictures for various indicators of achievement motivation, typically shows males to be much more oriented toward competitive effort than females are.
Males and females also differ in their propensity to take risks (Kogan & Wallach 1964). Generally speaking, women seem to be less given to taking risks than men, but the ingenious experiments these authors report show that the sex difference interacts in a complex way with variation in the nature of the task or situation and with the personality traits of anxiety and defensiveness.
Emotionality
There is considerable evidence from many sources that males and females differ in the strength of some of their emotional needs and the manner in which these are expressed. Some kinds of free responses in standardized situations suggest that these differences are related to body images and the differing requirements of sexual intercourse. Erikson (1951), for example, in the California guidance study asked the subjects to “construct an exciting movie scene” from materials provided. Boys typically produced high structures, ruins, and scenes suggesting sudden arrest of motion, whereas girls produced static, open enclosures, such as rooms. It is difficult to see how all of the differences reported in studies of this sort can be explained simply on the basis of differential treatment of boys and girls in home and community.
The emotional characteristic about which the largest amount of evidence has accumulated is aggression. Again and again, whether the subjects are preschool children, adult men and women, or any age group in between and whether the comparison is made on the basis of personality test scores, observations and ratings, or projective techniques, male groups score significantly higher on aggression than female groups. [SeeAggression.]
An equally consistent type of finding, although not resting on quite as substantial a body of research findings, is that females are more oriented toward other persons than males are. Sometimes this orientation toward people shows up as sensitivity and responsiveness, sometimes as passivity and dependence, sometimes as needs for succor-ance, nurturance, and affiliation. A finding that is perhaps related to these is that neuroticism seems to be more prevalent among females than among males, just as delinquency seems to be more prevalent among males than among females.
Summary of differences
Bennett and Cohen (1959), on the basis of a comprehensive study of their own, summarized the differences between the masculine and feminine approaches under five general principles. These fit in with most of the reported research on personality differences.
(1)Masculine thinking is of less intensity than feminine thinking.
(2)Masculine thinking is oriented more in terms of the self, whereas feminine thinking is oriented more in terms of the environment.
(3)Masculine thinking anticipates rewards and punishments determined more as a result of the adequacy or inadequacy of the self, whereas feminine thinking anticipates rewards and punishments determined more as a result of the friendship or hostility of the environment.
(4)Masculine thinking is associated more with desire for personal achievement, feminine thinking more with desire for social love or friendship.
(5)Masculine thinking finds value more in malevolent and hostile actions against a competitive society, whereas feminine thinking finds value more in freedom from restraint in a friendly and pleasant environment.
It must always be remembered that these sex differences are only group trends and the distributions for males and females on measures of any of them overlap to a considerable extent. As in the ability domain, there are marked individual differences within each sex. But the evidence suggests that the average difference between sex groups is somewhat greater for temperamental qualities than for abilities.
Masculinity-femininity scales
Because consistent sex differences show up on many kinds of items used in personality tests and because there are differences within each sex group in the responses individuals give to these items, many psychologists found it feasible to construct empirical masculinity-femininity (M-F) scales made up entirely of items that had been shown to differentiate the sexes. It was hoped that such scales would be useful for diagnostic and counseling purposes, as masculinity-femininity would seem to be a fundamental aspect of an individual’s personality.
The first and most comprehensive of these M-F scales was the Terman and Miles Attitude-Interest Analysis Blank (Terman & Miles 1936). The authors carried out an extensive research program using this test to determine how various special groups in the population distribute themselves with regard to the M-F variable. Among occupational groups, men who are athletes and engineers obtained the most masculine scores. Men who are journalists, artists, and clergymen were least masculine. Women who are domestic employees were most feminine; women who are athletes and doctors were least feminine. The age trends differed for the two sexes. Eighth-grade girls were the most feminine of all the groups tested, eleventh-grade boys the most masculine. With advancing age, males tended to become a little more feminine, females a little more masculine.
Since Terman and Miles began on this line of research, several other M-F scales have been constructed in the same manner. It is common practice for a multiscore test, like the Strong Vocational Interest Blank or the Minnesota Multiphasic Personality Inventory, to include an M-F key among the other scoring keys.
Critique of M-F scales
Although all of these scales have served a useful research purpose, they have not had as much practical value as had been expected. In general, they have not been of much use in diagnosing homosexuality. Terman and Miles in their original series of investigations showed that it was only the passive homosexuals (those who customarily played female roles in homosexual relationships) who made unusually feminine scores on the test, and since this effect was based to a large extent on the Interests sub-test, on which persons like artists and clergymen also tend to make feminine scores, one must be very cautious about drawing diagnostic conclusions from test performance. The same difficulty arises in connection with the later M-F scales. There are various factors that tend to produce deviant scores, so that one cannot assume they point to homosexuality.
The hope has been largely abandoned that M-F scales might play a useful part in counseling situations to help individuals find work and living situations that would fit in with their basic temperamental qualities. The difficulty is that masculinity-femininity is not unidimensional. Two persons with identical M-F scores may have obtained them from a very different combination of item responses. The emotional qualities that characteristically differentiate males from females show little or no correlation with masculine or feminine interests. It is quite possible for either a man or a woman to deviate in the direction of the opposite sex on one type of item and not on another. Several studies have produced unambiguous evidence of this lack of unidimensionality. One of the most recent (Barrows & Zuckerman 1960) showed that the correlations between three commonly used M-F scales were only about .3.
The organization of personality
A kind of research evidence that is being increasingly emphasized as psychologists attempt to incorporate what they know about sex differences into some coherent theoretical system is the repeated finding that variables are related differently to one another in male and female samples. Often the difference in the pattern of correlations between variables appears more impressive than the mean differences on any one variable. Such differences in correlational pattern have been identified in many areas of personality research, and in most cases the reasons for their existence have not yet been made clear. It is possible only to summarize these areas and in some cases suggest tentative explanations.
Discrepant patterns for the two sexes have repeatedly turned up in factor-analytic research on abilities and on personality traits. The only general conclusion one might draw from a miscellany of such studies is that perhaps females are less differentiated than males. There is a tendency for fewer factors to emerge from the data of females than from those of males.
The predictive validity of scholastic aptitude and achievement tests seems to be greater for females than for males (Seashore 1962). However, in longitudinal studies in which personality status at adulthood is to be predicted from childhood measurements of a particular trait (Kagan & Moss 1962), females are more predictable in some areas, males in others. The fact that the stability coefficients are higher on aggression for males and on passivity-dependence for females suggests the reasonable hypothesis that traits characterizing either sex as a whole are more stable within that sex group than are traits characterizing the other sex as a whole.
Different personality variables show significant relationships to general adjustment and popularity in males and females. Correlations often even turn out to be of opposite sign. The only general summary one might hazard of a number of studies in this area is that characteristics related to conformity of behavior and attitude bring more social rewards for females and that those related to initiative and independence bring more social rewards for males. Many specific findings, however, cannot be classified under this one statement.
The developmental antecedents of some traits, especially aggression and underachievement in school, are different for the two sexes. The growth curves differ in shape for these and several other characteristics that have been studied.
It is this last type of finding that has generated what appears to be the most fruitful research ideas for dealing with the whole mass of diverse material. What they suggest is that the development of male and female individuals follows a different course. Thus developmental concepts would seem to offer the most promising theoretical formulations.
Sex role concepts and behavior
Much of the developmental research of the 1950s and 1960s has focused on the learning of sex roles. Sociologists are primarily responsible for the formulation of the role concept, but it fits in well with several varieties of psychological personality theory. Historical as well as biological factors enter into the formulation of what the sex roles are in any given culture, but once these roles exist, the shaping of masculine and feminine personalities in growing children requires that each person learn in detail what these roles are and at the same time learn to prefer for himself what is considered appropriate for his sex. The purpose of much developmental research is to determine when and how such learning occurs and what factors influence it, both positively and negatively.
Various experimental techniques for finding out what young children consider the proper sex roles to be and how they relate themselves to them have been devised. These techniques include standardized doll-play situations, the telling of stories about pictures, and the choice of toys or activities for a hypothetical boy or girl. Investigators find with considerable consistency that children as young as five understand what is considered appropriate behavior for each sex in much the same way as adults do. Some researchers have worked with children as young as three and have shown that the same sex role concepts, not quite so precisely formulated, can be identified even at this early age. When children are asked to make choices or state preferences, interesting age trends appear. In general, increasing age, from the preschool years on, brings an increasing degree of preference for the characteristics and activities of one’s own sex, but there are irregularities and exceptions. Particularly at the early ages, girls show more preference for the activities of boys than boys show for those of girls (Brown 1956).
Identification
In trying to find out how such learning occurs, psychologists have made use of the concept of identification. It is difficult, if not impossible, to specify a precise meaning for this term, but most of the particular ways in which it has been used in research on sex differences have something to do with the way that an individual makes the standards and attitudes of his parents a part of himself. Some studies in this field have compared boys from families that are intact with boys from families in which the father is temporarily or permanently absent. Sex-typed behavior, such as aggression, develops at an earlier age in boys with fathers in the home than in boys from “father-absent” families; this suggests that sex roles are acquired through identification. The fact that most boys from “father-absent” homes eventually develop the sex-typed behavior, however, suggests that identification with the parent of the same sex is not the only way in which such roles can be acquired. Other studies have shown that preschool children are aware of what their parents expect of boys and girls and are influenced by these expectations. Father-son and mother-daughter similarities on tests of interests and values have been demonstrated.
In all of these types of investigation, many individual irregularities and exceptions to the prevailing trend of the evidence occur. The process through which a son learns from his father and a girl learns from her mother may not be identical. Most aspects of the mother’s sex role can be directly observed and imitated by her daughter, whereas the son does not have the same opportunity to observe and imitate his father’s actual behavior. Boys must thus proceed more on the basis of inference. Rapid social change complicates the problem of sex roles. What a child learns in his own home may not be completely appropriate to the situation that exists when he becomes an adult. These and many more considerations point to the conclusion that the learning of sex-appropriate behavior is an enormously complex process, as yet only incompletely understood.
General significance of research
As indicated in the introduction, the study of sex differences has been undertaken in order to answer important practical questions and in order to contribute to our knowledge of the ways in which many sorts of individual differences in personality come into existence. If the complicated interweaving of biological and social determiners can be unraveled from the fabric of sex differences in our culture, perhaps the pattern of other designs of still greater complexity can be traced. Both practical and theoretical purposes continue to influence the planning and conduct of research. In the middle 1960s, the theoretical purpose predominates. Research on sex differences is a strategic meeting ground for biologists, psychologists, anthropologists, and sociologists. An adequate theory, when one is achieved, must represent a synthesis of ideas from many sources.
Leona E. Tyler
[Other relevant material may be found inIdentity, PSYCHOSOCIAL; Role; Sexual Behavior.]
BIBLIOGRAPHY
Barrows, Gordon A.; and Zuckerman, Marvin 1960 Construct Validity of Three Masculinity-Femininity Tests. Journal of Consulting Psychology 24:441-445.
Bennett, Edward M.; and Cohen, Larry R. 1959 Men and Women: Personality Patterns and Contrasts. Genetic Psychology Monographs 59:101-155.
Brown, Daniel G. 1956 Sex-role Preference in Young Children. Psychological Monographs 70, no. 14.
Dunsdon, M. I.; and Roberts, J. A. F. 1957 A Study of the Performance of 2,000 Children on Four Vocabulary Tests. British Journal of Statistical Psychology 10:1-16.
Edgerton, Harold A.; and Britt, Steuart H. 1947 Technical Aspects of the Fourth Annual Science Talent Search. Educational and Psychological Measurement 7:3-21.
Erikson, E. H. 1951 Sex Differences in the Play Configurations of Preadolescents. American Journal of Orthopsychiatry 21:667-692.
Gaier, Eugene L.; and Collier, Mary J. 1960 The Latency-stage Story Preferences of American and Finnish Children. Child Development 31:431-451.
Kagan, Jerome; and Moss, Howard A. 1962 Birth to Maturity: A Study in Psychological Development. New York: Wiley.
Kogan, Nathan; and Wallach, Michael A. 1964 Risk Taking: A Study in Cognition and Personality. New York: Holt.
Mcclelland, David C. et al. 1953 The Achievement Motive. New York: Appleton.
Mcnemar, Quinn 1942 The Revision of the Stanford-Binet Scale. New York: Houghton.
Sandstrom, Carl I. 1953 Sex Differences in Localization and Orientation. Acta Psychologica 9:82-96.
Seashore, Harold G. 1962 Women Are More Predictable Than Men. Journal of Counseling Psychology 9: 261-270.
Strong, Edward K. Jr. 1943 Vocational Interests of Men and Women. Stanford (Calif.) Univ. Press.
Terman, Lewis M.; and Miles, Catherine C. 1936 Sex and Personality: Studies in Masculinity and Femininity. New York: McGraw-Hill.
Terman, Lewis M.; and Tyler, Leona E. 1954 Psychological Sex Differences. Pages 1064-1114 in Leonard Carmichael (editor), Manual of Child Psychology. 2d ed. New York: Wiley.
Tyler, Leona E. 1956 A Comparison of the Interests of English and American School Children. Journal of Genetic Psychology 88:175-181.
Witkin, Herman A. et al. 1954 Personality Through Perception. New York: Harper.