TE科2年生における欠点科目数の因果分析

構造方程式モデリングによるアプローチ

抄録

The causal relationship between the number of failing subjects on an exam and each subject were analyzed by means of structural equation modeling (SEM). The data used were the results of the first semester final examinations of sophomore belonging to the Department of Information, Communication and Electronic Engineering of Kumamoto National College of Technology. Since these data provide unsuitable solution, a stepwise method was applied to extract four subjects that affect the number of failing subjects. These subjects are mathematics, physics, computer engineering and engineering practice, which are the base of science and engineering. I also determined the number of latent variables by means of scree plots and dendrograms. Then I propose a causal model in which one latent variables behind the four subjects affect the number of failing subjects. The lavaan package on R was used to apply the SEM. From the goodness-of-fit indices, it was found that the model explained well the causal relationship related to the number of failing subjects.