Structural Equation Modeling

Structural equation modeling (SEM) is a data analysis technique used to specify and test a theoretical model of relations among variables. SEMs use an elaborate system of terms and notation to define and document various parts of the model. There are several different types of SEMs, including confirmatory factor analysis, path analysis, and latent growth modeling, as well as variants of those types. For each type of SEM, the degree to which the specified model fits the observed data is tested using multiple statistics. SEM is a flexible method for modeling relations among variables in life-span human development. This entry introduces SEM, describes the three major types of SEMs, and describes the current practices for testing the fit of SEMs.