Structural Equation Modeling
Structural Equation Modeling can be thought of as combination of factor analysis, path analysis, and regression analysis. It is a confirmatory cross-sectional linear statistical technique and it allows researchers to test and determine the validity of a certain model.
There are many variables which cannot be measured directly and they are called ‘latent variables’. Latent variables are very common in behavioral science domain (such as psychology) and they represent abstract concept that cannot be measured directly such as “intelligence” or “attitude”. These abstract concepts are measured using observed or manifest variables. Structural Equation Modeling allows you to model the relationship between these abstract concepts.
Structural equation modelling uses a structure of the covariance matrix of the abstract concepts (or latent variables). Once the model’s parameters have been estimated, the resulting implied covariance matrix (from the specified model) is compared to an empirical covariance matrix (from the data). The structural model is considered a plausible explanation for relations between the measures, once the two matrices are consistent with one another.