Age Differences in Primary Mental Ability Factors

T T IS well known that the intercorrelations •* • among intellectual tests are generally posi-tive. This fact provides the logical, scientific justification for the common practice of adding together scores obtained from diverse kinds of intellectual tasks to get a single score to repre-sent, operationally, a concept of general intelli-gence. Omnibus tests based upon this practice —tests such as the Stanford-Binet, Army Alpha, WAIS, Otis, etc.—have frequently been used to study the relationship between aging and change of intellectual ability. But as Cattell (1950; 1957) has emphasized, the fact of posi-tive manifold among the intercorrelations for a set of variables is not, in itself, sufficient evi-dence to support an hypothesis that the varia-bles represent a functional unity. And if the diverse subtests used in omnibus tests of intelli-gence do not form a functional unity, then studies of age differences in the scores obtained with these tests can be quite misleading. Factor analytic research has amply estab-lished the fact that the diverse tasks used in omnibus tests of intelligence involve several distinct ability functions, as represented by