Modeling the students' mathematics achievement in Malaysia

Abstract

The main goal of this research is to assess the efficiency of Structural Equation Modeling (SEM) and Normal Practice methods. This is because SEM is said to have a better estimation and prediction when multiple response items were involved in constructs. The efficiency of both methods was compared by using Absolute, Incremental and Parsimonious fitness indexes computed by AMOS graphic software. From the result of this research, it is found that SEM method achieved all fitness indexes required level while Normal Practice method failed to achieve Absolute and Parsimonious fit. As a conclusion, SEM method is proven to be more efficient in making estimation and prediction than the Normal Practice method.