Studying correlates and predictors of longitudinal change using structural equation modeling

Abstract

This paper is concerned with the study of correlates and predictors of change in a multiwave design. A general structural modeling approach is discussed, which allows estimation of theoretically and empirically relevant interrelationship indexes between growth or decline in longitudinally assessed psychological constructs and additional variables. Several classical test theory-based structural models are discussed. The models permit consistent and efficient estimation of, and tests about, the degree of covariation between change in one or more repeatedly measured latent dimensions and other variables, such as studied or presumed correlates of growth or decline in the longitudinally observed constructs. These models are useful in developmental studies with multiple assessment points, in which variables that are correlated with, and can be used to predict, change in measured abilities in repeatedly assessed psychological characteristics are to be identified. The approach is illustrated with data from a cognitive intervention study of aged adults (Baltes, Dittmann-Kohli, & Kliegl, 1986). Index terms: correlates and predictors of growth or decline, longitudinal research design, measurement of change, multiple assessments, structural equation modeling.