Studying correlates and predictors of longitudinal change using structural equation modeling

This paper is concerned with the study ofcorrelates and predictors of change in a multiwavedesign. A general structural modeling approach isdiscussed, which allows estimation of theoreticallyand empirically relevant interrelationship indexesbetween growth or decline in longitudinallyassessed psychological constructs and additionalvariables. Several classical test theory-based structuralmodels are discussed. The models permitconsistent and efficient estimation of, and testsabout, the degree of covariation between change inone or more repeatedly measured latent dimensionsand other variables, such as studied or presumedcorrelates of growth or decline in the longitudinallyobserved constructs. These models are usefulin developmental studies with multiple assessmentpoints, in which variables that are correlated with,and can be used to predict, change in measuredabilities in repeatedly assessed psychologicalcharacteristics are to be identified. The approachis illustrated with data from a cognitive interventionstudy of aged adults (Baltes, Dittmann-Kohli,& Kliegl, 1986). Index terms: correlates and predictorsof growth or decline, longitudinal researchdesign, measurement of change, multiple assessments,structural equation modeling.