Abstract

Structural equation model, which is widely used to describe the relationship between latent variables, can be judged by its goodness of fit. The test for statistical significance and the effect size index for practical significance of model fit use dichotomous and continuous interpretation approach to evaluate the usefulness of the model, respectively. However, despite the fact that the level of fit is represented on a continuum for practical significance, the calculated effect size index is interpreted dichotomously by using the guideline as an absolute standard. The present study discusses the process of assessing the practical significance of fit in terms of the effect size index and the correct use of guidelines so that researchers evaluating the fit of a model can interpret the level of fit on a continuum. We begin with a brief discussion of the importance of assessing statistical significance using test, and then define the concept of effect size in the context of structural equation models. We then introduce the different types of goodness of fit effect size indices and describe the characteristics of the guidelines used to interpret them. Finally, we provide examples of appropriate guidelines for interpreting calculated effect size index values on a continuum and discuss examples of incorrect model evaluation when continuity is not reflected, as well as the correct interpretation of models with marginal fit.