Fit Estimation in Structural Equation Modeling - Synthesis of Related Statistics

Abstract

Structural Equation Modeling (SEM) as a multivariate data analysis technique is considered sophisticated one and becoming popular among researchers. Literature is not consistent on one of the important matter of SEM which is related with the determination of adequacy of model fit. In view of this shortcoming, the paper aims at accumulating opinions and notions of researchers which are evident from literature regarding fit of SEM model on the basis of indices of fit. Afterward, it contently analyzes previous studies to give a new outlook to the topic. A criterion is proposed in the paper, based on which academicians and researchers can determine the level of fit for their estimated models. The estimated models may not always be perfect and even may not be acceptable as per the guidelines provided in literature. But the paper further proposes that any research model cannot be rejected just because the data which is analyzed does not fit it perfectly. Therefore, both researchers and practitioners are encouraged to contribute on the said aspect. To further develop the topic and enhancement in the research field of SEM, the studies with strong theoretical background even with poor fitting models should be published. Let the new studies do their job of examining, whether it is the research model that is not appropriate or the problem is associated only with the poor quality and less reliable data.

Keywords: Structural Equation Modeling, Measurement Model, Structural Model, Fit Indices, Model Fit