VALIDITY

Interpretation of test scores ultimately involves predictions about a subject's behavior in a specified situation. If a test is an accurate predictor, it is said to have good validity. Before validity can be demonstrated, a test must first yield consistent, reliable measurements. In addition to reliability, psychologists recognize three main types of validity.

A test has content validity if the sample of items in the test is representative of all the relevant items that might have been used. Words included in a spelling test, for example, should cover a wide range of difficulty.

Criterion-related validity refers to a test's accuracy in specifying a future or concurrent outcome. For example, an art-aptitude test has predictive validity if high scores are achieved by those who later do well in art school. The concurrent validity of a new intelligence test may be demonstrated if its scores correlate closely with those of an already well-established test.

Construct validity is generally determined by investigating what psychological traits or qualities a test measures; that is, by demonstrating that certain patterns of human behavior account to some degree for performance on the test. A test measuring the trait “need for achievement,” for instance, might be shown to predict that high scorers work more independently, persist longer on problem-solving tasks, and do better in competitive situations than low scorers.

                                          

Reliability refers to the consistency of test scores. A reliable test yields the same or close to the same score for a person each time it is administered. In addition, alternate forms of the test should produce similar results. By these criteria, modern intelligence tests are highly reliable. In fact, intelligence tests are the most reliable of all psychological tests.

Validity is the extent to which a test predicts what it is designed to predict. Intelligence tests were designed to predict school achievement, and they do that better than they do anything else. For example, IQ scores of elementary school students correlate moderately with their class grades and highly with achievement test scores. IQ tests also predict well the number of years of education that a person completes. The SAT is somewhat less predictive of academic performance in college. Educators note that success in school depends on many other factors besides intelligence, including encouragement from parents and peers, interest, and motivation.

Intelligence tests also correlate with measures of accomplishment other than academic success, such as occupational status, income, job performance, and other measures of vocational success. However, IQ scores do not predict occupational success as well as they predict academic success. Twenty-five percent or less of the individual differences in occupational success are due to IQ. Therefore, a substantial portion of the variability in occupational success—75 percent or more—is due to factors other than intelligence.

Validity also refers to the degree to which a test measures what it is supposed to measure. A valid intelligence test should measure intelligence and not some other capability. However, making a valid intelligence test is not a straightforward task because there is little consensus on a precise definition of intelligence. Lacking such a consensus, test makers usually evaluate validity by determining whether test performance correlates with performance on some other measure assumed to require intelligence, such as achievement in school.