Abstract
This article illustrates a specification error problem in a simple structural model. A quadratic relationship, described by a linear structural model with a latent variable, is shown to have less predictive validity than a simple manifest variable regression model. The use of simpler analyses that require fewer structural assumptions is advocated as a preliminary analysis to latent variable structural modeling. A discussion of the role of statistical models in the detection of causal relationships is also provided.
Recommended Citation
Pohlmann, John T.
(1993)
"Insidious Structural Errors in Latent Variable Models,"
Mid-Western Educational Researcher: Vol. 6:
Iss.
3, Article 9.
Available at:
https://scholarworks.bgsu.edu/mwer/vol6/iss3/9
COinS
