Structural equation models (SEM) and their application in education

Abstract

Structural equation modeling (SEM) is a statistical methodology that takes a confirmatory (hypothesis testing) approach to the analysis of a structural theory that relates to some phenomenon. Typically, this theory represents causal processes, which generate observations in multiple variables (Bentler, 1988). The term structural equation model conveys two important aspects of the procedure: (a) that the causal processes under study are represented by a series of structural equations, regression; and (b) that these structural relationships can be modeled graphically to allow a clearer conceptualization of the theory under study (Byrne, 2010). These structural equation models were born from the need to provide greater flexibility to regression models. They are less restrictive than regression models because they allow measurement errors to be included in both the criterion (dependent) variables and the predictor (independent) variables. The SEM could be thought of as several factor analysis models that allow for direct and indirect effects between factors. The great advantage of this type of models is that they allow proposing the type and direction of the relationships that are expected to be found between the various variables contained in it, in order to subsequently estimate the parameters that are specified by the proposed theoretical relationships. In this article, the characteristics and phases of structural equation models are exposed, such as the construction stages, the interpretation of results, the graphical representation and the types of variables used. The possible use of these models in educational research at the university level is also addressed.