ABSTRACT
In the social sciences it has become customary to denote by structural equation modeling
(SEM) the area that focuses on estimating models involving systems of equations that
include latent variables, measurement errors, multiple indicators, and so on. Most often
such models are estimated by specifying the covariance structure implied by the model
and estimating the model parameters by minimizing a discrepancy function between the
observed covariance matrix and the sample covariance matrix. Thus, covariance structure
modeling is a special case of structural equation modeling. For a good overview of
structural equation modeling the reader may consult Bollen (1989), Browne and
Arminger (1995), Bentler and Dudgeon (1996), and Yuan and Bentler (1997).