STA 4107/5107
Chapter 10:
Introduction to Structural Equation Modeling
March 27, 2007
1
Key Terms
Please review and learn these terms. SEM is extremely jargonladen. I even recommend for this
topic that you create flashcards or vocabulary lists that you can memorize and refer to while
studying. Below, new and jargon terms will be in italics. Please look them up if you need to. I
will be defining these terms to some extent, but life will be easier for all of us if you have some
exposure to them before lecture. My admonition to read the chapter before lecture is really crucial
for this material.
Below are some of the more important terms you will need to be familiar with for this lecture:
Causal inference
A dependence relationship of two or more variables in which the researcher
clearly specifies that one or more variables is the
cause
of an outcome represented by at least
one other variable.
Causation
The principle by which cause and effect are established between two variables.
It
requires a sufficient degree of association (covariance), that the cause occurs before the effect,
and that no other reasonable causes for the outcome are present. Strictly, causation is almost
never proven, but in practice strong theoretical support can make empirical estimation of
causation possible.
Communality
The total amount of variance that a
measured
variable has in common with the
constructs
upon which it loads.
Good measurement practice suggests that each measured
variable should load on only one construct.
Hence, it can be thought of as the variance
explained in a measured variable by the construct. In
confirmatory factor analysis
(CFA), it
is referred to as the squared multiple correlation for a measured variable.
Confirmatory analysis
The use of a multivariate technique to test (confirm) a prespecified rela
tionship. It is the opposite of
exploratory analysis
.
Construct
An unobservable or
latent
concept that the researcher can define in conceptual terms
but that cannot be directly measured (i.e. there is no single measurement that can be made
that will totally and perfectly quantify the concept), or cannot be measured without error.
1
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Endogenous constructs
The
latent
, multivariate equivalents to dependent variables. It is repre
sented by a variate of dependent variables.
Exogenous constructs
The
latent
, multivariate equivalents to independent variables. They are
constructs determined by factors outside the model.
Fixed parameter
A parameter that has a value specified by the researcher. Most often the value
specified is zero, indicating no relationship, although in some instances a nonzero value can
be specified.
Free parameter
A parameter estimated by the structural equation program to represent the
strength of a specified relationship. These parameters may occur in the
measurement model
(most often denoting loadings of
indicators
to constructs), as well as in the
structural model
(relationships among constructs).
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '08
 Staff
 Linear Regression, Regression Analysis, measurement model

Click to edit the document details