The Effect of the Difference in the Estimation Method and the Sample Size on Goodness of Fit Indexes in the Path Analysis Model as one of the modeling patterns in the Structural Equation in Psychological and Educational Research

Abstract

The study aimed to compare the goodness of fit indexes for the path analysis model resulting from the use of two methods of estimation: Maximum likelihood method (ML) and generalized least squares method (GLS) through different samples sizes (n = 200, 400, 600, 800, 1000) according to the rule (n + 200) using computer- generated data that follow a normal distribution.

     The results of the study showed that there are no fundamental differences between the quality indicators that are identical to the path analysis model resulting from the use of two methods to estimate the proposed path analysis model.

     The results also showed the efficiency of the goodness of fit indexes for the path analysis model and their fall in the ideal range in both methods through the different samples. The results of the study indicated the importance of having a sample size at least (n= 200) when using the path analysis model.

     The study recommended the use of several of ideal goodness of fit indexes when using the path analysis method as one of the structural equation modeling (SEM) patterns in psychological and educational research.