Explores the four-step procedure for testing structural equation models and outlines some problems with the approach advocated by L. Hayduk and D. Glaser (2000) and S. Mulaik and R. Milsap (2000). Questions the idea that there is a "correct" number of constructs for a given phenomenon. (SLD)

Focuses on the four-step method (four nested models) of structural equation modeling advocated by S. Mulaik (1997, 1998), discussing the limitations of the approach and considering the tests and criteria to be used in moving among the four steps. (SLD)

Defends the four-step approach to structural equation modeling based on testing sequences of models and points out misunderstandings of opponents of the approach. The four-step approach allows the separation of respective constraints within a structural equation model. (SLD)

Neither the four-step model nor the one-step procedure can actually tell whether the researcher has the right number of factors in structural equation modeling. In fact, for reasons discussed, a simple formulaic approach to the correct specification of models does not yet exist. (SLD)

Replies to commentaries on the four-step approach to structural equation modeling, pointing out the strengths and weaknesses of each argument and ultimately concluding that the four-step model is subject to criticisms that can be addressed to factor analysis as well. (SLD)

Quality of life (QOL) has become an important concept for health care. As QOL is a multidimensional concept that is best evaluated by a number of latent constructs, it is well recognized that latent variable models, such as exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) are useful tools for analyzing QOL data. Recently,…

Developed a Bayesian approach for a general multigroup nonlinear factor analysis model that simultaneously obtains joint Bayesian estimates of the factor scores and the structural parameters subjected to some constraints across different groups. (SLD)

Examines some positions in various arguments related to the proper number of factors and proper number of steps when using structural equation models. Defines the issue in estimating structural equation models as a problem of specifying a model appropriately based on theoretical concerns and then diagnosing ills in the model as well as possible.…

Considers models incorporating principal components from the perspectives of structural-equation modeling. These models include the following: (1) the principal-component analysis of patterned matrices; (2) multiple analysis of variance based on principal components; and (3) multigroup principal-components analysis. Discusses fitting these models…

This article reviews several basic statistical tools needed for modeling data with sampling weights that are implemented in Mplus Version 3. These tools are illustrated in simulation studies for several latent variable models including factor analysis with continuous and categorical indicators, latent class analysis, and growth models. The…

Discusses two criticisms raised by L. Hayduk and D. Glaser of the most commonly used point estimate of the Root Mean Square Error (RMSEA) and points out misconceptions in their discussion. Although there are apparent flaws in their arguments, the RMSEA is open to question for several other reasons. (SLD)

Presents an approach (designed for beginners) to using PROC CALIS, a Statistical Analysis System (SAS) procedure, to perform path analyses using observed variables. The approach begins with the development of a figure to illustrate the researcher's theoretical model, and then converts the figure into a PROC CALIS program. (SLD)

Discusses issues related to model evaluation in structural equation modeling. Supports nested model comparisons via sequential chi-square difference tests as consistent with the four-step approach to model evaluation when models of the factor analytic simultaneous equation type are entertained. (Author/SLD)

Demonstrates empirically a structural equation modeling technique for group comparison of reliability and validity. Data, which are from a study of 495 mothers' attitudes toward pregnancy, have a one-factor measurement model and three sets of subpopulation comparisons. (SLD)

Two methods, direct maximum likelihood (ML) and the expectation maximization (EM) algorithm, can be used to obtain ML parameter estimates for structural equation models with missing data (MD). Although the 2 methods frequently produce identical parameter estimates, it may be easier to satisfy missing at random assumptions using EM. However, no…