Goodness of fit as a single factor structural equation model
The assessment of model data fit in structural equation modeling is a complicated and rapidly expanding area of research in the social sciences. Numerous fit indices have been proposed by researchers which attempt to capture the degree to which a structural equation model agrees with the observed data. Numerous studies have shown that different fit indices are affected to different extents by such things as degree of non-normality of the data, estimation method employed, and mis-specification within the model. Despite these differing properties, new students to structural equation modeling are often told that the fit indices will all tend to agree with each other in their assessment of the model data fit. This study investigated whether or not this agreement manifests itself in a single factor model in which multiple fit indices load on a latent factor that could be thought of as the construct good fit. The results of this study showed that the one factor model was an over-simplification. Rather than loading on a single factor, the fit indices included in this study were affected by two factors. The two factors can be thought of as absolute discrepancy and comparative discrepancy. Additionally, results of this study showed that the observed fit indices were affected differentially by the two latent factors. Suggestions for further research were made, including one based on a somewhat surprising result found with respect to the goodness of fit index and other fit indices derived from it.
|Year of publication:||
|Authors:||Gullen, James Andrew|
Wayne State University
|Type of publication:||Other|
ETD Collection for Wayne State University
Saved in favorites
Similar items by subject
Find similar items by using search terms and synonyms from our Thesaurus for Economics (STW).