Applications of structural equation modeling to psychological research

This paper's aim is to discuss some applications of structural equation mod eling (SEM) to psychological research. SEM integrates factor and path analy sis techniques, making possible the analysis of latent variables (theoretic al constructs) and the estimation of relations between those variables, sim ultaneously in a single model. In addition, SEM allows for the estimation o f goodness of fit between specific theoretical postulates of a psychologica l model, and data used to test such model. Because of these features, struc tural models allow for: a) building latent variables from observed indicato rs, b) testing the psychometric properties (reliability, validity) of instr uments expected to assess such latent variables, c) estimating hypotheses r egarding relations between latent variables, and d) determining the pertine nce (goodness of fit) of a theoretical model representing relations among b ehavioral variables. These applications are illustrated by using data from a study on environmental-conservation behavior. It is concluded that the us e of structural equation modeling will positively affect psychological rese arch in the close future.