Multidimensional Scaling And Structural Equation Modeling: A Comparison Of Multivariate Techniques For Theory Testing

A "causal modeling" technique, structural equation modeling, is applied in conjunction with a recently developed multidimensional scaling technique, CMDA (Borg and Lingoes, 1980). CMDA (Constrained Monotone Distance Analysis) allows one to place theoretically-derived constraints on a multidimensional scaling solution, and to evaluate both measurement and theory through a series of "regional hypotheses" (Lingoes, 1980). These alternative methods are applied in an evaluation of a social-psychological model derived from the Likert (1961, 1967) theory of organizational behavior. The results demonstrate the utility of CMDA for assessing discriminant validity in a multivariate context, and discuss the extent to which this sort of discriminant validity is a necessary condition for the accurate interpretation of a latent variable structural equation model.