Analysis of adult age differences on the Raven's Advanced Progressive Matrices tes

The purpose of the current project was to examine the nature of the age-related differences on the Raven's Advanced Progressive Matrices (APM). In the first of two studies, a rational analysis of the APM was conducted in which four components were hypothesized to be involved in correctly solving the APM problems. These included decomposition, rule-identification, rule-application, (involving a one-rule spatial transformation), and rule-coordination. In addition, performance on perceptual speed was examined as a potential contributor to performance on the hypothesized components. Two tests presumed to measure each hypothesized component and three tests presumed to measure perceptual speed were presented to 165 young adults. Correlational and confirmatory analyses indicated that the measures of Rule Application and Rule Coordination were reasonable estimates of the constructs and, thus, were included in Study 2. However, the measures of Decomposition were not strongly related, nor sufficiently distinguishable from perceptual speed to warrant inclusion in Study 2. In addition there was only a weak relationship between the measures of Rule Identification and so one of the measures was replaced for Study 2. The two tasks of the three remaining hypothesized components and the three measures of perceptual speed were presented to 183 subjects between the ages of 21 and 83. Study 2 was designed to examine the influence of each of the hypothesized components on the total and the age-related variance on the APM. In addition the role of perceptual speed was examined in relation to performance on both the APM and the hypothesized components. Hierarchical regression analyses indicated that although all of the hypothesized components accounted for a significant amount of the variance on the APM (approximately 50% each), the measures of Rule Application accounted for a the largest unique proportion of both the total and the age-related variance on the APM. In addition, the measures of Rule Application accounted for age-related variance on the APM beyond that accounted for by perceptual speed and working memory. It is suggested that the measures of Rule Application share a meaningful relationship with the APM that is independent of the other two components.