The psychometric approach to intelligence emphasizes people’s performance on standardized aptitude tests. Aptitude tests predict people’s future ability to acquire skills or knowledge. Achievement tests, on the other hand, measure skills and knowledge that people have already learned.
Intelligence tests can be given individually or to groups of people. The best-known individual intelligence tests are the Binet-Simon scale, the Stanford-Binet Intelligence Scale, and the Wechsler Adult Intelligence Scale.
Alfred Binet and his colleague Theodore Simon devised this general test of mental ability in 1905, and it was revised in 1908 and 1911. The test yielded scores in terms of mental age. Mental age is the chronological age that typically corresponds to a particular level of performance.
Example: A ten-year-old child whose score indicates a mental age of twelve performed like a typical twelve-year-old.
In 1916, Lewis Terman and his colleagues at Stanford University created the Stanford-Binet Intelligence Scale by expanding and revising the Binet-Simon scale. The Stanford-Binet yielded scores in terms of intelligence quotients. The intelligence quotient (IQ) is the mental age divided by the chronological age and multiplied by 100. IQ scores allowed children of different ages to be compared.
Example: A ten-year-old whose performance resembles that of a typical twelve-year-old has an IQ of 120 (12 divided by 10 times 100).
There are two problems with the intelligence quotient approach:
The Stanford-Binet was revised in 1937, 1960, 1973, and 1986.
David Wechsler published the first test for assessing intelligence in adults in 1939. The Wechsler Adult Intelligence Scale contains many items that assess nonverbal reasoning ability and therefore depends less on verbal ability that does the Stanford-Binet. It also provides separate scores of verbal intelligence and nonverbal or performance intelligence, as well as a score that indicates overall intelligence.
The term intelligence quotient, or IQ, is also used to describe the score on the Wechsler test. However, the Wechsler test presented scores based on a normal distribution of data rather than the intelligence quotient. The normal distribution is a symmetrical bell-shaped curve that represents how characteristics like IQ are distributed in a large population. In this scoring system, the mean IQ score is set at 100, and the standard deviation is set at 15. The test is constructed so that about two-thirds of people tested (68 percent) will score within one standard deviation of the mean, or between 85 and 115.
On the Wechsler test, the IQ score reflects where a person falls in the normal distribution of IQ scores. Therefore, this score, like the original Stanford-Binet IQ score, is a relative score, indicating how the test taker’s score compares to the scores of other people. Most current intelligence tests, including the revised versions of the Stanford-Binet, now have scoring systems based on the normal distribution. About 95 percent of the population will score between 70 and 130 (within two standard deviations from the mean), and about 99.7 percent of the population will score between 55 and 145 (within three standard deviations from the mean).
Individual intelligence tests can be given only by specially trained psychologists. Such tests are expensive and time-consuming to administer, and so educational institutions often use tests that can be given to a group of people at the same time. Commonly used group intelligence tests include the Otis-Lennon School Ability Test and the Lorge-Thorndike Intelligence Test.
Some researchers have suggested that biological indices such as reaction time and perceptual speed relate to intelligence as measured by IQ tests:
Many psychologists believe that cultural bias can affect intelligence tests, for the following reasons:
Some characteristics of IQ tests are standardization, norms, percentile scores, standardization samples, reliability, and validity.
Intelligence tests are standardized, which means that uniform procedures are used when administering and scoring the tests. Standardization helps to ensure that people taking a particular test all do so under the same conditions. Standardization also allows test takers to be compared, since it increases the likelihood that any difference in scores between test-takers is due to ability rather than the testing environment. The SAT and ACT are two examples of standardized tests.
Researchers use norms when scoring the tests. Norms provide information about how a person’s test score compares with the scores of other test takers. Norms allow raw test scores to be converted into percentile scores. A percentile score indicates the percentage of people who achieved the same as or less than a particular score. For example, if someone answered twenty items correctly on a thirty-item vocabulary test, he receives a raw score of 20. He consults the test norms and finds that a raw score of 20 corresponds with a percentile score of 90. This means that he scored the same as or higher than 90 percent of people who took the same test.
Psychologists come up with norms by giving a test to a standardization sample. A standardization sample is a large group of people that is representative of the entire population of potential test takers.
Most intelligence tests have good reliability. Reliability is a test’s ability to yield the same results when the test is administered at different times to the same group of people. For more on reliability, see page 14.
Validity is a test’s ability to measure what it is supposed to measure. For more on validity, see page 14. Although intelligence tests cannot be considered good measures of general intelligence or general mental ability, they are reasonably valid indicators of the type of intelligence that enables good academic performance.