New Directions for Statistical Methods in Higher Education Research in Japan

A Review and Reflection on Experiences in Past 10 Years

Abstract

　 In this paper, we reviewed articles applying quantitative analysis to higher education field published in two academic journals and/or an institutional journal over the past 10 years in Japan.

　 As a result, we discovered the following points; 1） we need to avoid incomplete information related to analysis, inappropriate application of analytical techniques, excessive interpretation, etc., 2） we need to solve problems as above by applying correctly new methods such as Generalized Liner Model, Zero-Inflated Model, Log-Linear Model, Decision Tree Analysis, Fixed Effect, Random Effect, Mixed Effect Model and Mixed Trajectory Model in near the future, 3） we need to consider the advantages of Supporting Information Service which allow us to complement information that we could not include in article.

　 Then, we indicated the importance to apply advanced methods like not only the methods described above but state-of-the-art ones as Text Mining, Bayes statistics, Machine Learning. Finally, we pointed out that all members of Higher Education Research Community are largely responsible for improvement of the quality and the excellence of quantitative analysis in higher education research.