The purpose of this project was to extend previous research (Babcock, 1994) which examined the influence of three hypothesized components, processing speed, and working memory on the individual- and age-related differences on the Raven's Advanced Progressive Matrices (APM) test. The results of that study suggested that although all three components are important to performance on the APM, rule-application tasks seemed to hold the most promise in accounting for age-related variance on the APM. The current project investigated two issues: (a) whether measures used to represent constructs important to APM performance were invariant for young and old, and (b) whether the influence of the hypothesized components on performance on the APM is similar for young and old adults. Participants included 134 young and 134 older adults. Results suggested that the measures represent similar constructs for both age groups. In addition, the structural model of performance on the APM was similar for young and old. Results are discussed in terms of previous research which found rule application to be promising in accounting for age differences on the APM.