Submodels and applications in covariance structure analysis

In recent years, social sciences are beginning to draw upon a new generation of multivariate analysis methods. In this paper, covariance structure analysis which is one of these methods, is discussed as a general method for many other multivariate analysis methods using the covariance structure expression by Toyoda (1990). The advantage of this method is that a priori knowledge (for instance, theoretical prediction, research design and empirical findings) can be incorporated into empirical analysis. Some worthy applications in education and psychology are introduced.