The construction and use of instruments for measuring INTELLIGENCE. The first attempt to measure intelligence scientifically was made by Francis Galton (1822–1911) in 1884 using tests of sensory discrimination and reaction time, but the tests turned out not to correlate with one another or with measures of scholastic performance. In 1905 Alfred Binet (1857–1911) and Theodore Simon (1873–1961) devised a test to detect and classify mentally retarded children in France. It was translated and introduced into the United States in 1916 as the STANFORD-BINET SCALE, which became the prototype of most subsequent tests of intelligence and the foundation of modern intelligence testing.
The Binet-Simon scale was designed to measure children’s ability to follow instructions, exercise judgement, and solve reasoning problems. Relying on the fact that cognitive ability generally increases as children grow older, Binet & Simon selected 54 items according to how well they discriminated in practice between children of different ages. This led to the establishment of test norms and the introduction of the concept of mental age. For example, if a child passed just those items that average 6-year-olds in the standardization sample could pass, the child was said to be functioning at a mental age of 6 years. The German psychologist William Stern (1871–1938) pointed out in 1912 that MENTAL AGE (MA) indicates intelligence only in relation to actual or CHRONOLOGICAL AGE (CA), and he introduced the concept of the INTELLIGENCE QUOTIENT, which he defined as the ratio MA/ CA. The American psychologist Lewis Terman (1877–1956) later introduced the abbreviation IQ (intelligence quotient) and suggested multiplying Stern’s ratio by 100 to convert it to a percentage, yielding the familiar formula IQ = MA/CA×100, according to which an IQ of 100 is average by definition.
Although Terman’s formula continues to appear in elementary textbooks, it is no longer used in psychometric practice. It was first abandoned by the American psychometrician David Wechsler (1896–1981) in 1939, largely because it leads to absurdities when applied to adults. Cognitive ability stops increasing around 18 years of age, yet a 40-year-old who performs at the level of an average 20-year-old could be said to have an IQ of 20/40×100 = 50, suggestive of severe mental retardation. To avoid the problems of the old formula, Wechsler introduced a purely statistical definition of IQ according to which IQ is a normally distributed variable with a mean or average of 100 and a standard deviation (a statistical measure of variability) of 15, and test scores are converted to this scale. Because of the known properties of the normal distribution, this means that 68% of IQ scores fall between 85 and 115, 95% between 70 and 130, and so on. This statistical definition of IQ has been generally adopted by test constructors since World War II.
A revised form of the Stanford-Binet scale is still used by researchers and educational psychologists. There are three different Wechsler scales for use with different age groups: the WECHSLER ADULT INTELLIGENCE SCALE (WAIS), the Wechsler Intelligence Scale for Children, and the Wechsler Pre-school and Primary Scale of Intelligence. Each of these scales has to be administered individually by a trained tester, and each yields separate scores for verbal IQ, performance IQ, and full-scale IQ. The British Ability Scales, first published in 1979, are a set of 23 tests designed to measure a wide diversity of mental abilities, yielding separate scores for visual IQ, verbal IQ and general IQ. RAVEN’S PROGRESSIVE MATRICES, first published in 1938 and revised several times since then, is a non-verbal test designed to measure abstract reasoning ability through the use of geometric diagrams. There are many other intelligence tests in common use, including various multiple-choice tests that can be administered to large groups of respondents at the same time. Interest has been growing since the late 1980s in biological measures of intelligence based on EVENT-RELATED POTENTIALS and in processing speed as a possible basis for intelligence, as first suggested by Galton.
Scores on different IQ tests correlate highly with one another, partly because a new test is not considered valid unless it correlates with existing tests. The most impressive finding to emerge from decades of research into intelligence is that different tests and subtests correlate with one another no matter how different they are in content. Most contemporary psychologists interpret this fact as evidence for a single common factor, which was labelled g (for general intelligence) by the British psychologist Charles Spearman (1863–1945) in 1927. A rival interpretation, first suggested by the American psychologist Louis Leon Thurstone (1887–1955) in 1938, focuses on the fact that the correlations are not perfect and that certain types of test items usually correlate more highly with one another than with other types. Thurstone believed that these clusters indicate the existence of seven independent primary mental abilities, which he labelled number, word fluency, verbal meaning, memory, reasoning, spatial perception, and perceptual speed.
ANDREW M.COLMAN
This is the complete article, containing 810 words
(approx. 3 pages at 300 words per page).
