Intelligence

Intelligence and intelligence tests are often at the heart of controversy. Some arguments concern the ethical and moral implications of, for example, selective breeding of bright children. Other arguments deal with the statistical basis of various conclusions such as whether tests are biased, or how much of intelligence is genetically determined. What one hears less often is discussion of the construct of intelligence itself: What is intelligence? How does it grow? How and why do people differ intellectually? Questions like these, along with many others, which are central to any discussion of intelligence and intelligence testing, are less often raised, much less answered.

Historical Roots of Intelligence Tests

Intelligence testing began as a more or less scientific pursuit into the nature of differences in human intellect. However, it soon acquired practical significance as a tool for predicting school achievement and selecting individuals for various educational programs. Sir Francis Galton's work in the late 1800s formed the background for much of the research and theory pursued during the twentieth century on the assessment of individual differences in intelligence.

Galton believed that all intelligent behavior was related to innate sensory ability but his attempts to empirically validate that assumption were largely unsuccessful. In France, Alfred Binet and Victor Henri (1896) criticized the approach advocated by Galton in England and James McKeen Cattell in the United States and argued that appropriate intelligence testing must include assessment of more complex mental processes, such as memory, attention, imagery, and comprehension. In 1904, Binet and Théodore Simon were commissioned by the French Minister of Public Instruction to develop a procedure to select children unable to benefit from regular public school instruction for placement in special educational programs. In 1905 Binet and Simon published an objective, standardized intelligence test based on the concepts developed earlier by Binet and Henri. The 1905 test consisted of thirty subtests of mental ability, including tests of digit span, object and body part identification, sentence memory, and so forth. Many of these subtests, with minor modifications, are included in the Stanford-Binet intelligence test of the early twenty-first century.

In 1908 and again in 1911, Binet and Simon published revised versions of their intelligence test. The revised tests distinguished intellectual abilities according to age norms, thus introducing the concept of mental age. The subtests were organized according to the age level at which they could be successfully performed by most children of normal intelligence. As a result, children could be characterized and compared in terms of their intellectual or mental age. The Binet and Simon intelligence test was widely adopted in Europe and in the United States. Lewis Terman of Stanford University developed the more extensive Stanford-Binet test in 1916. This test has been used extensively in several updated versions throughout the United States.

A major change in intelligence testing involved the development of intelligence tests that could be simultaneously administered to large groups. Group tests similar to the original Binet and Simon intelligence test were developed in Britain and the United States. During World War I, group-administered intelligence tests (the Army Alpha and Army Beta tests) were used in the United States to assess the abilities of recruits who could then be selected for various duties based on their performance. In England, from the 1940s to the 1960s, intelligence tests were administered to all children near the age of eleven years to select students for different classes of vocational training.

An enormous number of "mental" tests are available in the early twenty-first century and are typically divided into those involving group versus individual administration. Whereas IQ (the intelligence quotient) was originally reported as the ratio of an individual's mental age to chronological age multiplied by 100 (100 × MA/CA), IQ has long been based upon normative score distributions for particular age groups. All individual and group tests currently yield such deviation IQs where 100 typically represents the 50th percentile and 68 percent of all scores fall between 85 and 115.

Factor Theories of Intelligence

What is intelligence and what do these tests actually assess? Very early in the psychological study of intelligence, Charles Spearman (1904) sought to empirically determine the similarities and differences between various mental tests and school performance measures. He found that many seemingly diverse mental tests were strongly correlated with each other. This led him to postulate a general factor of intelligence, g, that all mental tests measure in common while simultaneously varying in how much the general factor contributes to a given test's performance. On the basis of correlational studies, Spearman argued that intelligence is composed of a general factor that is found in all intellectual functioning plus specific factors associated with the performance of specific tasks. Spearman's theoretical orientation and methods of analysis served as the foundation of all subsequent factor analytic theories of intelligence. Spearman (1927) later developed a more complex factor theory introducing more general "group factors" made up of related specific factors. However, he adhered to his main tenet that a common ability underlies all intellectual behavior. For lack of a better definition, he referred to this as a mental energy or force.

The concept that intelligence is characterized by a general underlying ability plus certain task-or domain-specific abilities constitutes the basis of several major theories of intelligence, including those offered by Cyril L. Burt (1949), Philip E. Vernon (1961), and Arthur Jensen (1998). Quite distinct from theories of intelligence that emphasize g are those that emphasize specific abilities that can be combined to form more general abilities. Lloyd L. Thurstone (1924, 1938) developed factor analytic techniques that first separate out specific or primary factors. Thurstone argued that these primary factors represent discrete intellectual abilities, and he developed distinct tests to measure them. Among the most important of Thurstone's primary mental abilities are verbal comprehension, word fluency, numerical ability, spatial relations, memory, reasoning, and perceptual speed.

Raymond Cattell (1963, 1971) attempted a rapprochement of the theories of Spearman and Thur-stone. In an attempt to produce a g factor, he combined Thurstone's primary factors to form secondary or higher-order factors. Cattell found two major types of higher order factors and three minor ones. The major factors were labeled gf and gc, for fluid and crystallized general intelligence. Cattell argued that the fluid intelligence factor represents an individual's basic biological capacity. Crystallized intelligence represents the types of abilities required for most school activities. Cattell labeled the minor general factors gv, gr, and gs for visual abilities, memory retrieval, and performance speed, respectively. Cattell's initial theory has been substantially extended by individuals such as John L. Horn (1979,1985).

The most recent psychometric research supports a hierarchical model of intellect generally in accord with the outlines of the Cattell-Horn theory. At the top of the hierarchy is g and under are broad group factors such as gf and gc. Below these broad group factors are more specific or narrow ability factors. The majority of intelligence tests focus on providing overall estimates of g (or gf and gc) since this maximizes the prediction of performance differences among people in other intellectual tasks and situations, including performance in school.

Alternative Theoretical Perspectives

The hierarchical model of human intelligence that has evolved from the psychometric or measurement approach is not the only influential perspective on human intellect. A second view of intelligence, that provided by developmental psychology, stems from the theory of intellectual development proposed by the Swiss psychologist Jean Piaget. This tradition is a rich source of information on the growth and development of intellect. A third view on intelligence, the information-processing or cognitive perspective, is an outgrowth of work in cognitive psychology since the 1970s. It provides elaborate descriptions and theories of the specific mental activities and representations that comprise intellectual functioning. The three perspectives are similar with regard to the general skills and activities that each associates with "being or becoming intelligent." For all three, reasoning and problem-solving skills are the principal components of intelligence. A second area of overlap among the three involves adaptability as an aspect of intelligence.

The differences and separate contributions of the three perspectives to an understanding of human intellect also stand out. The emphasis on individual differences within the psychometric tradition is certainly relevant to any complete understanding of intelligence. A theory of intelligence should take into account similarities and differences among individuals in their cognitive skills and performance capabilities. However, a theory of human intellect based solely on patterns of differences among individuals cannot capture all of intellectual functioning unless there is little that is general and similar in intellectual performance.

In contrast, the developmental tradition emphasizes similarities in intellectual growth and the importance of organism—environment interactions. By considering the nature of changes that occur in cognition and the mechanisms and conditions responsible, one can better understand human intellectual growth and its relationship to the environment. This requires, however, that one focus not just on commonalities in the general course of cognitive growth, but consider how individuals differ in the specifics of their intellectual growth. Such a developmental-differential emphasis seems necessary for a theory to have adequate breadth and to move the study of intelligence away from a static, normative view, where intelligence changes little over development, to a more dynamic view that encompasses developmental change in absolute levels of cognitive power.

Finally, the cognitive perspective helps to define the scope of a theory of intelligence by further emphasizing the dynamics of cognition, through its concentration on precise theories of the knowledge and processes that allow individuals to perform intellectual tasks. Psychometric and developmental theories typically give little heed to these processes, yet they are necessary for a theory of intelligence to make precise, testable predictions about intellectual performance.

No theory developed within any of the three perspectives addresses all of the important elements and issues mentioned above. This includes the more recent and rather broad theories such as Howard Gardner's multiple intelligences theory (1983, 1999) and Robert J. Sternberg's triarchic theory (1985). Both theories represent an interesting blending of psychometric, developmental, and cognitive perspectives.

Uses and Abuses of Tests

Above it was noted that testing was developed in response to pragmatic concerns regarding educational selection and placement. The use of intelligence tests for educational selection and placement proliferated during the decades from the 1930s through 1960s as group tests for children became readily available. Since the early 1980s, however, general intelligence testing has declined in public educational institutions. One reason for diminished used of such tests is a trend away from homogeneous grouping of students and attendant educational tracking. A second reason is that achievement rather than aptitude testing has become increasingly popular. Not surprisingly, such tests tend to be better predictors of subsequent achievement than aptitude or intelligence tests. Even so, it is an established fact that measures of general intelligence obtained in childhood yield a moderate 0.50 correlation with school grades. They also correlate about 0.55 with the number of years of education that individuals complete.

Intelligence and aptitude tests continue to be used with great frequency in military, personnel-selection, and clinical settings. There are also two major uses of intelligence tests within educational settings. One of these is for the assessment of mental retardation and learning disabilities, a use of tests reminiscent of the original reason for development of the Binet and Simon scales in the early 1900s. The second major use is at the postsecondary level. College entrance is frequently based upon performance on measures such as the SAT, first adopted in the United States by the College Entrance Examination Board in 1937. Performance on the SAT, together with high school grades, is the basis for admission to many American colleges and universities. The ostensible basis for using SAT scores is that they moderately predict freshman grade point average–precisely what they were originally designed to do. However, considerable debate has arisen about the legitimacy and value of continued use of SAT scores for college admission decisions.

Throughout the history of the testing movement, dating back to the early 1900s and extending to the early twenty-first century, there has been controversy concerning the (mis) use of test results. One of the earliest such debates was between Lewis Terman, who helped develop the revised Stanford-Binet and other tests, and the journalist Walter Lippman. A frequent issue in debates about the uses and abuses of intelligence tests in society is that of bias. It is often argued that most standardized intelligence tests have differential validity for various racial, ethnic, and socioeconomic groups. Since the tests emphasize verbal skills and knowledge that are part of Western schooling, they are presumed to be unfair tests of the cognitive abilities of other groups. As a response to such arguments, attempts have been made to develop culture-fair or culture-free tests. The issue of bias in mental testing is beyond this brief review and Arthur R. Jensen (1980; 1981) can be consulted for highly detailed treatments of this topic. Evidence in the 1990s suggests that no simple form of bias in either the content or form of intelligence tests accounts for the mean score differences typically observed between racial and ethnic groups.

Factors Affecting Test Scores

Much of the research on intelligence has focused on specific factors affecting test scores. This includes research focused on environmental versus genetic contributions to IQ scores, related issues such as race differences in IQ, and overall population trends in IQ.

One of the most extensively studied and hotly debated topics in the study of intelligence is the contribution of heredity and environment to individual differences in test scores. Given a trait such as measured intelligence on which individuals vary, it is inevitable for people to ask what fraction is associated with differences in their genotypes (the so-called heritability of the trait) as well as what fraction is associated with differences in environmental experience. There is a long history of sentiment and speculation with regard to this issue. It has also proven difficult to answer this question in a scientifically credible way, in large measure due to the conceptual and statistical complexity of separating out the respective contributions of heredity and environment. Adding to the complexity is the need to obtain test score data from people who have varying kinship and genetic relationships, including identical and fraternal twins, siblings, and adoptive children with their biological and adoptive parents. Nonetheless, evidence has slowly been gathered that heritability is sizeable and that it varies across populations. For IQ, heritability is markedly lower for children, about 0.45, than for adults where it is about0.75. This means that with age, differences in test scores increasingly reflect differences in the genotype and in individual life experience rather than differences in the families within which they were raised. The factors underlying this shift as well as the mechanism by which genes contribute to individual differences in IQ scores are largely unknown. The same can be said for understanding environmental contributions to those differences. A common misconception is that traits like IQ with high heritability mean that the results are immutable, that the environment has little or no impact, or that learning is not involved. This is wrong since heritable traits like vocabulary size are known to depend on learning and environmental factors.

Perhaps no topic is more controversial than that of race differences in IQ, especially since it is so often tied up with debates about genetics and environmental influences. It is an established fact that there are significant differences between racial and ethnic groups in their average scores on standardized tests of intelligence. In the United States, the typical difference between Caucasians and African Americans is 15 points or one standard deviation. A difference of this magnitude has been observed for quite a long period of time with little evidence that the difference has declined despite significant evidence that across the world IQ scores have risen substantially over the last fifty years. The latter phenomenon is known as the "Flynn Effect," and it, like so many other phenomena associated with test scores, begs for an adequate explanation.

There is a tendency to interpret racial and ethnic differences in mean IQ scores as being determined by genetic factors since, as noted above, IQ scores in general have a fairly high level of heritability and the level of heritability seems to be about the same in different racial and ethnic groups. There is, however, no logical basis on which to attribute the mean difference between racial groups to either genetic or environmental factors. As one group of researchers stated, "In short, no adequate explanation of the differential between the IQ means of Blacks and Whites is presently available" (Neisser et al., p. 97).

Age and Intelligence

Although most intelligence tests are targeted for school-age populations, there are instruments developed for younger age groups. Such tests emphasize the assessment of perceptual and motor abilities. Unfortunately, measures of infant and pre-school intelligence tend to correlate poorly with intelligence tests administered during the school years. However, there appears to be a high degree of stability in the IQ scores obtained in the early primary grades and IQ scores obtained at the high school level and beyond. Often this is misinterpreted as indicating that an individual's intelligence does not change as a function of schooling or other environmental factors. What such results actually indicate is that an individual's score relative to his or her age group remains fairly constant. In an absolute sense, an individual of age 16 can solve considerably more difficult items and problems than an individual of age 8. Comparing IQ scores obtained at different ages is akin to comparing apples and oranges since the composition of tests changes markedly over age levels.

Research has also studied changes in IQ following early adulthood. A frequent conclusion from research examining age groups ranging from 21 to 60 and beyond is that there is an age-related general decline in intellectual functioning. However, there are serious problems with many such studies since they involve cross-sectional rather than longitudinal contrasts. In those cases where longitudinal data are available, it is less obvious that intelligence declines with age. John L. Horn and Cattell (1967) presented data indicating a possible differential decline in crystallized and fluid intelligence measures. Crystallized intelligence measures focus on verbal skills and knowledge whereas fluid intelligence measures focus on reasoning and problem solving with visual and geometric stimuli. The latter also often place an emphasis on performance speed. Fluid intelligence measures tend to show substantial declines as a function of age, whereas crystallized intelligence measures often show little or no decline until after age 65. Research in the 1990s based on combinations of longitudinal and cross-sectional samples supports the conclusion that there are age-related declines in intelligence, which seem to vary with the type of skill measured, and that the declines are often substantial in the period from age 65 to 80.

Future Outlook

After more than 100 years of theory and research on the nature and measurement of intelligence there is much that researchers know but even more that they don't understand. Still lacking is any agreed upon definition of intelligence and many of the empirical findings regarding intelligence test scores remain a puzzle. In their summary paper "Intelligence: Knowns and Unknowns," Ulric Neisser and colleagues (1996) stated:

In a field where so many issues are unresolved and so many questions unanswered, the confident tone that has characterized most of the debate on these topics is clearly out of place. The study of intelligence does not need politicized assertions and recriminations; it needs self restraint, reflection, and a great deal more research. The questions that remain are socially as well as scientifically important. There is no reason to think them unanswerable, but finding the answers will require a shared and sustained effort as well as the commitment of substantial scientific resources. (p. 97)