Providing education raises the difficult issue of individual differences, in which the effects of lecture methods, details, and materials vary according to learners’ characteristics (aptitudes and personalities). If this individual variability is not deliberated about regarding the measurement of educational effectiveness, only the improvement of “students’ average value” will be considered. In such a case, the educational effectiveness of all learners cannot be reasonably enhanced.

Maltby and Macaskill (2007) is detailed for studies on individual differences. Ikemoto, Seki and Tsubaki (2005) examined a questionnaire survey design to analyze the relationship between demand for qualitative improvement of high school education and characteristics of students. Tsubaki and Wakabayashi (2008) studied learning support based on the individual differences of learners through a Decision Tree. Tsubaki, Kakuta and Murata (2009) analyzed the individual differences of learning method of mathematics using constrained categorical conjoint analysis. Placing more importance on educational effects, Tsubaki, Tsuchida, Kimura and Watanabe (2009) proposed an analysis of effectiveness that was conducted in consideration of the individual differences of learners using a Bayesian Network. See Castillo, Gutierrez, and Hadi(1997), Jensen(2001), Pearl and Russell(2003), Korb and Nicholson(2004), Ben-Gal(2007) and Motomura (2007) in detail about Bayesian Network.

In this paper, we construct an analytical system of educational effects considering the learners’ individual differences based on the analysis proposed by Tsubaki, Tsuchida, Kimura and Watanabe(2009).

In this study, we obtain data on student's learning activity, extracurricular activities, school events, and characteristics, etc. by questionnaire in a high school. Moreover, we obtain test scores by examining student's scholastic attainments. Then, we analyze the structure of the student life by the factorial analysis. Based on the factor scores of the result of this factorial analysis, students are classified by clustering (Ward method). Next, we analyze whether the factors that influence scholastic attainments are different or not in each class by the structural equation modeling. In addition, we analyze the differences among classes in detail, how the variables obtained by structural equation modeling effect to scholastic attainments in each class. Consequently, we analyze how the educative effect of the school is different according to the student type.

These analyses are combined, and then we construct an analytical system for analyzing educational effects that takes individual differences into account using Rexsel.

As application example, we analyze data of above high school students. We think about items that becomes the center of the high school life. Then, 55 research questions of questioners are categorized to 14 items. These items are “Willingness to learn,” “Result of the class (score),” “Test,” ”Class policy in the future,” “Teaching evaluation,” “Wishing reason,” “Desire,” “Working hard,” “Characteristics,” “Interpersonal relationship,” “Extracurricular activities,” “School events,” ”Equipment and environment” and “Satisfaction rating of the entire high school life.”

We categorize 1499 students who go to a high school by their responses for questioners, obtained 4 types of students Next, we analyze the factors effect the students’ scores using structural equation modeling. Further, we show the conditional probabilities obtaining the scores, given levels of the important factor (study approach).

[1] Maltby, J. Day, L. & Macaskill, A. (2007): Personality, Individual Differences and Intelligence. London: Pearson Education [2] Ikemoto, K., Seki, H. and Tsubaki, M.(2005): “ Questionnaire Survey Design and Model Construction for Analyzing Relationship between Demand for Qualitative Improvement of High School Education and Characteristics of Students,” The Japanese Journal of Behaviormetrics, 32, pp.1-19. [3] Tsubaki,M. and Wakabayashi,S.(2008): “Proposal of SRM and Segmentation of Students by Learning Type”, Journal of Japan Industrial Management Association, 59, pp.269–281. [4] Tsubaki,M., Kakuta,T. and Murata,S. (2009)：Constrained Categorical Conjoint Analysis, New Trends in Psychometrics IMPS 2007 Conference Volume，pp.481-490. [5]M. Tsubaki, Y. Tsuchida, K.Kimura and M. Watanabe：” Analysis of the Educational Effectiveness Considering Individual differences using Bayesian Network, Proceedings of European Conference on Educational Research (ECER) 2009, (2009) [6] Castillo,E., Gutierrez,J.M. and Hadi,A.S.(1997): Expert Systems and Probabilic Network Models.New York;Springer-Verlag. [7] Jensen,F.V.(2001): Bayesian Networks and Decision Graphs, Springer. [8] Pearl,J. and Russell,S.(2003): Bayesian Networks, in Arbib,M.A.(editor),Handbook of Brain Theory and Neural Networks, pages 157-160,Cambridge, MA:MIT Press. [9] Korb,K.B. and Nicholson,A.E.(2004):Bayesian Artificial Intelligence. CRC Press. [10] Ben-Gal,I.(2007)：Bayrsian Networks, in Ruggeri,F., Kenett,R. and Faltin,F.(editors), Encyclopedia of Statistics in Quality and Reliability, John Wiley & Sons. [11] Motomura,Y. (2007): “New Development in Basic and Applied Bayesian Network”, Journal of the Japanese Society for Artificial Intelligence, 22, 302-305. [12] Anderson,J.C.(1987):”Structural equation models in the social and behavioral sciences: Model building. Child Development, 58, 49-64.