A Comparison Between Ordinary Least Square (OLS) And Structural Equation Modeling (SEM) Methods In Estimating The Influencial Factors Of 8th Grades Student’s Mathematics Achievement In Malaysia

This research demonstrates the application of Structural Equation Modeling (SEM) method in order to obtain the best fit model for a more efficient and accurate inter-relationship among variables findings and interpretation. For the purpose of this study, secondary data of Trends in International Mathematics and Science Study (TIMSS) had been used. The data were distributed by using two stage stratified cluster sampling technique and involving 5733 eighth grades students in Malaysia. A Confirmatory Factor Analysis (CFA), Discriminant Validity and Path Analysis had been conducted to obtain the best fit model of SEM. At the end of the study, the best fit model was then compared to the Ordinary Least Square (OLS) model method. From the Chi-Square value, it is found that the model of SEM method is much better compared to OLS model method based on its fitness and accuracy. KeywordsMathematics Achievement, Mediation, Ordinary Least Square, Structural Equation Modeling, TIMSS ——————————  ——————————