A cross-sequential study of age changes in cognitive behavior

Previous cross-sectional and longitudinal studies of age changes over the adult life span have reported contradictory age gradients. The apparent contradiction was assessed by means of a new research design, called the crosssequential method, which involves the repeated measurement of members of a cross-sectional sample. The SRA Primary Mental Abilities Test (PMA) and Schaie's Test of Behavioral Rigidity (TBR) were administered to a stratifiedrandom sample of 500 5s with quotas of 25 men and 25 women in each 5-yr. age interval from 20 to 70 yrs. 7 yr. later all 5s who could be located were contacted and 302 5s were retested. Significant cross-sectional age changes were found for all variables studied, but longitudinal age changes occurred for all cohorts only for those variables where response speed was of importance. Analysis of the comparative age gradients suggests that age changes over time within a given individual appear to be much smaller than differences between cohorts and that the steep textbook age gradients represent no more than the effects of increased environmental opportunity and/or genetic changes in the species. Further implications with respect to revisions in current thinking on adult age changes are discussed. One of the pervading problems troubling the developmental psychologist who is interested in studying age changes over the adult life span has been the consistent reporting of contradictory age gradients found as the result of cross-section al and longitudinal inquiries. Many cross-sectional studies report peak performance in the early twenties or thirties with steep decrement gradients thereafter (cf. Horn & Cattell, 1966; Jones & Conrad, 1933; Schaie, 1958; Schaie, Rosenthai, & Perlman, 19S3). Most longitudinal studies, on the other hand, report no decrement at all. In fact, slight gains in performance are recorded at least into the midfifties (Bayley & Oden, 19SS; Owens, 1953). It has been argued that these contradictory findings can be accounted for by systematic sample attrition in the longitudinal studies, which tends to eliminate more subjects of low ability. It has been observed that none of the longitudinal studies has yet reached the sixties and seventies, the age range where the greatest decrement has been noted in the cross-sec