Abstract
In recent years, the environment of the entrance examination of Japanese universities is drastically changing. On the other hand, the prediction method based on the chronological order analysis with regard to an entrance rate (number of the new students/ number of the successful candidates) may be chosen in each university deciding a pass boundary line. However, such the chronological order analysis about the result variables is lacking prediction power. Especially, when the change of the environment is intense, the risk that causes a large error must be high. In this report, we will try to propose the prediction model based on the expression of a structure of causality that is comparatively stable against yearly change. The proposed model will combine both of the macroscopic viewpoints and microscopic one. The former pays attention to yearly change of the entrance rate, and the latter pays attention to judgment about whether or not each examinee enters.
