Structural models of the relations between age and measures of cognitive functioning

Alternative models of the pattern of relations among age and variables representing distinct cognitive factors were examined in data from a study involving 206 adults between IS and 84 years of age. A model postulating that the variables each had independent age-related influences did not provide a good fit to the data, but models with a single age-related influence were surprisingly accurate at reproducing the age correlations on the variables. The best fit among the models examined was obtained with a hierarchical model in which the age-related influences operate only at highest level, on what is common to all variables. Because only age-sensitive variables were included in the analyses and because reasoning variables were closely related to the highest level factor, the hierarchical pa;tern is consistent with the fluid intelligence (Gf) model of Hem and Cattell. Additional analyses on data from this and other studies revealed that there was a strong positive relation between a variable's loading on the first principal component representing what all the variables have in common and the absolute value of the correlation of the variable with age. These results imply that broad or general factors play an important role in the age-related declines reported in many different types of cognitive tests. (C) 2001 Elsevier Science Inc. All rights reserved.