文档介绍：15
I I I I Ill I Illlll
STRUCTURAL EQUATION
MODELING: USES AND ISSUES
LISABETH F. DILALLA
School of Medicine, Southern Illinois University, Carbondale, llinois
Structural equation modeling (SEM) has e a popular research tool
in the social sciences, including psychology, management, economics, sociol-
ogy, political science, marketing, and education, over the past two to three
decades. Its strengths include simultaneous assessment of various types of
relations among variables and the ability to rigorously examine pare
similarities among and differences between two or more groups of study
participants. However, one of its major limitations is the ease with which
researchers can misinterpret their results when anxious to "prove" the
validity of a model or to attempt to assess causality in the relation between
two or more variables when the research design does not allow for such
conclusions. These will be elaborated below. First, however, it is essential
to understand exactly what SEM is in order to better understand when it
is most useful to apply it as a statistical tool. The purpose of this chapter
is to provide an overview of some of the possible uses of this technique
and to discuss some of the important assumptions and necessary conditions
for using SEM and for interpreting the results.
Handbook of Applied Multivariate Statistics and Mathematical Modeling
Copyright 2000 by Academic Press. All rights of reproduction in any form reserved. 439
440 LISABETH F. DILALLA
I. DEFINING STRUCTURAL EQUATION MODELING
Latent variable analysis (Bentler, 1980) involves the study of "hidden"
variables that are not measured directly but that are estimated by
variables that can be measured. Latent variable analysis includes such
techniques as factor analysis, path analysis, and SEM. Factor analysis
(see also Cudeck, chapter 10, and Hoyle, chapter 16, this volume)
involves assessing a latent factor that is operationa