Causal Models with Unmeasured Variables: An Introduction to Lisrel

Abstract

Recent methodological advances (Joreskog and Sorbom, 1978) now allow the estimation of causal models which incorporate structural relationships among latent variables, and confirmatory factor procedures to estimate latent variables and the measurement properties of their manifest indicators. The examination of structural effects among theoretical variables not directly measurable offers great promise for developing and testing theory in educational research. This paper presents an introduction to such models.