ABSTRACT
Structural equation modeling (SEM) is an advanced modeling approach that allows estimating latent variables as the common variance from multiple measures. SEM holds promise to account for measurement error and method biases, which allows one to get more accurate estimates of constructs, people's standing on constructs (i.e., individual differences), and associations between constructs.
Path analysis is similar to multiple regression but also allows inclusion of multiple dependent variables in the same model. SEM is path analysis but with latent variables. A SEM model consists of a measurement model and a structural model. The measurement model is a confirmatory factor analysis (CFA) model that identifies how many latent factors are estimated, and which items load onto which latent factor. The structural component of a SEM model includes the regression paths that specify the hypothesized causal relations among the latent variables. SEM is CFA, but it adds regression paths that specify hypothesized causal relations between the latent variables, which is called the structural component of the model.
