On Sensitivity of Structural Equation Modeling to Latent Relation Misspecifications

Abstract

By giving counterexamples where the covariance matrix provides crucial information about consequential model misspecifications, this article cautions researchers about overinterpreting Rogosa and Willett's (1985, p. 104) conclusion that "the covariance matrix is a severe summary of the longitudinal data, one that may discard crucial information about growth."