Tests of Intelligence:
The first tests designed to measure intelligence were developed by Sir Francis Galton. But for present use Galton’s tests are useless.
The intelligence test as we know it today was formulated by the French psychologist Alfred Binet (1857-1911). The French government asked Binet to devise a test that would detect those children too slow intellectually to profit from regular schooling. He assumed that intelligence should be measured by tasks requiring reasoning and problem solving, rather than perceptual-motor skills. In collaboration with Theodore Simon (1873-1961), another French psychologist, Binet published a scale in 1905, which he revised in 1908 and again in 1911. These Binet scales are the direct predecessors of contemporary intelligence tests.
Binet devised a scale of units of mental age. Average mental-age (MA) scores correspond to chronological age (CA), that is, to the age determined form the date of birth. A bright child’s MA is above his CA; a dull child has an MA below his CA. The mental-age scale is easily interpreted by teachers and others who deal with children differing in mental ability.
There are two chief ways to find items for which success is uninfluenced by special training. One way is to choose novel items with which an untaught child has as good a chance to succeed as one who has been taught at home or in school.
Many of the items on an intelligence test of the Binet type assume general familiarity. A vocabulary test, for example, appears in almost all the scales. Familiarity with the standard language of the test is necessarily assumed.
It is not enough to look at an item and to decide that it requires intelligence to answer it successfully. Some “tricky” or “clever” items turn out to be poor because of the successes or failures that occur through guessing. More pedestrian items, such as matters of common information, sometimes turn out to be most useful. These are “fair” items if all have had a chance to learn the answers.
These two requirements for an acceptable item (increase in percentage passing with age and I correlation with total score) reflect both validity and reliability. The first requirement is an indirect way I of guaranteeing validity, being based on the inference that what we mean by intelligence should [distinguish an older child from a younger one; the second requirement is a guarantee of reliability I through internal consistency of the measures.
Contemporary Binet Tests:
The tests originally developed by Binet underwent several revisions in this country, the first by Goddard in 1911. For many years the best-known and most widely used revision was that made by Terman at 1 Stanford University in 1916, commonly referred to as the Stanford-Binet. The test was revised in 1937, 1 1960, and 1972.
In the Binet tests, an item is age-graded at the level at which a substantial majority of the children I pass it. The present Stanford-Binet has six items of varied content assigned to each year, each item j when passed earning a score of two months of mental age.
Aptitude tests designed to predict performance over a broad range of abilities are called intelligence tests. Other aptitude tests measure more specific abilities; mechanical aptitude tests measure various types of eye-hand coordination; musical aptitude tests measure discrimination of pitch, rhythm, and other aspects of musical sensitivity that are predictive of musical performance with training; and clerical aptitude tests measure efficiency at number-checking and other skills that have been found to be predictive of an individual’s later achievement as an office clerk.
Many aptitude tests have been constructed to predict success in specific jobs or vocations. Since the Second World War the armed I forces have devised tests to select pilots, radio technicians, submarine crews, and many other I specialists.
One way to obtain information on specific kinds of abilities, rather than a single mental-age score, is to separate the items into more than one group and to score the groups separately. The Wechsler Adult Intelligence Scale (described in Table) and the Wechsler Intelligence Scale for Children use items similar to those in the Binet tests, but they divide the total test into two parts-a verbal scale and a performance scale-according to the content of the items.