Construction and Application of Log-linear Models to Assess Academic Performance

Log-Linear Models (LLMs) are important techniques used in categorical data analysis. Though there are some available published work about LLMs, the explanation of model building process and the theoretical background are not adequate. Furthermore, research about the application of the LLM theory and the selection procedure of the best model are handful. Therefore, this manuscript aims to fill that vacuum. At first, the construction of LLM and Hierarchical Log-Linear Models (HLLMs), a branch of LLMs are discussed in connection with both 2×2 and 2×2 ×2 contingency tables. Secondly, an application is presented to analyze the collected data set about the academic performance of elementary students. The manuscript also discusses the criteria to select the best model that fits the collected data