Latent Variable Structural Equation Models: LISREL in Special Education Research

A conceptual understanding of a broad range of research methods is important for those beginning or continuing a career in special education research. This article presents an overview of nonexperimental causal methods, with a particular focus on latent variable structural equation modeling through the LISREL computer program. An extended example is used to present LISREL as an extension of structural equations analysis (path analysis) and as a method of reducing the effects of error in research. Suggested steps are presented for conducting a LISREL analysis, and other uses of the method are presented. It is argued that nonexperimental research is important in special education, and that causal methods such as LISREL are especially appropriate for the analysis on nonexperimental data.