The School of Science and Engineering of Kinki University conducts a survey with a questionnaire given to all students at the end of every semester. The aim of this survey is to improve the teaching system by assessing the level, atmosphere of each class and enthusiasm of instructors from students' viewpoints. In other words, the aim is to enhance educative effects by establishing meaningful relations between students and instructors. To verify the effectiveness of the current survey with a questionnaire, this paper presents the statistical discussion using multivariate analysis. The data used here are the evaluations by sophomores, who studied the Analysis for Statistical Data at the department of industrial management. The analysis adopted here are Multiple Regression Analysis and Principal Component Analysis. The resulted multiple regression model has clarified the relationship between the methods of teaching and students evaluation. Also since the usage of the Principal Component Analysis has identified how students' evaluations had been reflected on the questionnaires results, the useful information to improve the teaching systems could be obtained.