Componential analysis as a way to investigate age-related changes in information processing

Abstract

The purpose of this paper is to alert researchers to the methods and utility of componential analysis as a means to examine age-related changes within information processing models of cognition. This analysis allows the researcher to determine which hypothesized information processing components are significant and the amount of variance shared with the dependent variable(s). Individual differences are investigated by modeling data at the individual subject level. Unstandardized regression weights are correlated with performance on a number of standardized ability tests to determine which components contribute to which abilities. The procedure combines complementary aspects of information processing analysis and psychometric analysis. Componential analysis is illustrated in this study with data from 60 individuals, aged 20 to 79, who solved verbal forced-choice analogies of the form A: B :: C : (D1 or D2). Solution time and error rate data were modeled using a regression model developed by Sternberg (1977).