A tutorial on structural equation modeling

Structural equation modeling (SEM) is a flexible statistical modeling technique that allows researchers to test hypotheses among observed and latent variables. This article intends to introduce basic concepts and procedures of SEM. Specifically, common terminologies and notations, modeling procedures, and practical issues in applications of SEM are discussed. A concrete example is used to illustrate major steps of SEM that include model specification, estimation, and evaluation. Special attention is dedicated to considerations in the choice of estimation methods particularly related to categorical data problems and the process of decision making.