The Detection and Correction of Specification Errors in Structural Equation Models

Saris, Willem E.; Satorra, Albert; Sörbom, Dag

doi:10.2307/271030

Cited by 166 publications

(152 citation statements)

References 13 publications

Supporting

Mentioning

149

Contrasting

Unclassified

Order By: Relevance

“…This in turn builds on older measures of Modification Indices and Expected Parameter Change (Jöreskog & Sörbom 1986;Saris et al 1987;Saris et al 2009) where the focus is on…”

Section: Sensitivity Analyses -Observed Vs Synthetic Datamentioning

confidence: 99%

Nonequivalence of measurement in latent variable modeling of multigroup data: A sensitivity analysis.

Kuha¹,

Moustaki²

2015

Psychological Methods

View full text Add to dashboard Cite

MEASUREMENT NON-EQUIVALENCE: A SENSITIVITY ANALYSIS 2In studies of multiple groups of respondents, such as cross-national surveys and cross-cultural assessments in psychological or educational testing, an important methodological consideration is the comparability or "equivalence" of measurement across the groups. Ideally full equivalence would hold, but very often it does not. If non-equivalence of measurement is ignored when it is present, substantively interesting comparisons between the groups may become distorted. We consider this question in multigroup latent variable modeling of multiple-item scales, specifically latent trait models for categorical items. We use numerical sensitivity analyses to examine the nature and magnitude of the distortions in different circumstances, and the factors which affect them. The results suggest that estimates of multigroup latent variable models can be sensitive to assumptions about measurement, in that non-equivalence of measurement does not need to be extreme before ignoring it may substantially affect cross-group comparisons. We also discuss the implications of such findings on the analysis of large comparative studies.

show abstract

“…This in turn builds on older measures of Modification Indices and Expected Parameter Change (Jöreskog & Sörbom 1986;Saris et al 1987;Saris et al 2009) where the focus is on…”

Section: Sensitivity Analyses -Observed Vs Synthetic Datamentioning

confidence: 99%

Nonequivalence of measurement in latent variable modeling of multigroup data: A sensitivity analysis.

Kuha¹,

Moustaki²

2015

Psychological Methods

View full text Add to dashboard Cite

show abstract

“…Saris et al [31] argued against the reliance on χ 2 test statistics and fit indices for model evaluation because they are not only affected by the degree of misspecification but also by the incidental characteristics of the model. Alternatively, they proposed to use MI along with EPCs.…”

Section: Sem-based Fit Indicesmentioning

confidence: 99%

“…EPC has been suggested to be used in conjunction with MI to detect model misspecification (see e.g., [30]). Several variations of EPC have been proposed: the unstandardized expected parameter change (EPC; [31]), which provides the estimated value that a given fixed parameter would have if it were freely estimated in the model; the partially standardized EPC [32], and the fully standardized EPC (SEPC; [33]), referred to as "Std YX E.P.C." in the Mplus package (2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016).…”

Section: Sem-based Fit Indicesmentioning

confidence: 99%

Specifying Turning Point in Piecewise Growth Curve Models: Challenges and Solutions

Ning

Luo

2017

Front. Appl. Math. Stat.

View full text Add to dashboard Cite

Piecewise growth curve model (PGCM) is often used when the underlying growth process is not linear and is hypothesized to consist of phasic developments connected by turning points (or knots or change points). When fitting a PGCM, the conventional practice is to specify turning points a priori. However, the true turning points are often unknown and misspecifications of turning points may occur. The study examined the consequences of turning point misspecifications on growth parameter estimates and evaluated the performance of commonly used fit indices in detecting model misspecification due to mis-specified locations of turning points. In addition, this study introduced and evaluated a newly developed PGCM which allows unknown turning points to be freely estimated. The study found that there are severe consequences of turning point misspecification. Commonly used model fit indices have low power in detecting turning point misspecification. On the other hand, the newly developed PGCM with freely estimated unknown turning point performs well in general.

show abstract

“…Therefore, we think that for all practical purposes the CU model can be seen as acceptable for this data set, and we display the actual estimates in Table 3. We consequently attribute the statistical rejection of the model to the probable high power of the chi-square test, caused by the fact that the standardized trait loadings are quite high (between .76 and .90) and that the sample size is relatively large (Saris & Satorra, 1988;Saris et al, 1987). 2.…”

Section: Testing Additive and Multiplicative Cu Modelsmentioning

confidence: 99%

“…In addition, both test statistics and the now-popular fit indexes may be sensitive to different characteristics of the model. The sensitivity of test statistics to those characteristics can be taken into account, although it is rather complex to do so, as has been shown by Satorra (1988, 1993); Saris, Satorra, and Sörbom (1987); Saris and Stronkhorst (1984); and Satorra and Saris (1985). Without going into details, we suggest using the following procedure:…”

Section: Testing Additive and Multiplicative Cu Modelsmentioning

confidence: 99%

Testing Nested Additive, Multiplicative, and General Multitrait-Multimethod Models

Coenders¹,

Saris²

2000

Structural Equation Modeling: A Multidisciplinary Journal

Self Cite

View full text Add to dashboard Cite

The article gives alternatives to Campbell and O'Connell's (1967) definitions of additive and multiplicative method effects in multitrait-multimethod (MTMM) data. The alternative definitions can be formulated by means of constraints in the parameters of the correlated uniqueness (CU) model (Marsh, 1989), which is first reviewed. The definitions have 2 major advantages. First, they allow the researcher to test for additive and multiplicative method effects in a straightforward manner by simply testing the appropriate constraints. An illustration of these tests is given. Second, the alternative definitions are closely linked to other currently used models. The article shows that CU models with additive constraints are equivalent to constrained versions of the confirmatory factor analysis model for MTMM data (Althauser, Heberlein, & Scott, 1971;Werts & Linn, 1970). In addition, Coenders and Saris (1998) showed that, for designs with 3 methods, a CU model with multiplicative constraints is equivalent to the direct product model (Browne, 1984).Multitrait-multimethod (MTMM) designs (Campbell & Fiske, 1959) consist of multiple measures of a set of factors (traits) with the same set of measurement procedures (methods). So these designs include t × m measures, that is, the number of methods (m) times the number of traits (t). The differences between methods can be STRUCTURAL EQUATION MODELING, 7(2),

show abstract

The Detection and Correction of Specification Errors in Structural Equation Models

Cited by 166 publications

References 13 publications

Nonequivalence of measurement in latent variable modeling of multigroup data: A sensitivity analysis.

Nonequivalence of measurement in latent variable modeling of multigroup data: A sensitivity analysis.

Specifying Turning Point in Piecewise Growth Curve Models: Challenges and Solutions

Testing Nested Additive, Multiplicative, and General Multitrait-Multimethod Models

Contact Info

Product

Resources

About