Comparing English Education Research in Japan to International Trends:

A Text Mining Study

Abstract

This paper aims to identify the characteristics of English language teaching (ELT) research in Japan. For this purpose, this study analyses conference abstracts from five domestic ELT associations (CELES, JACET, JALT, JASELE and KATE), two international ELT associations (AsiaTEFL and TESOL International Association) and Association for Japanese Language Education. Instead of interpretive coding techniques commonly used in previous studies, this study employs text mining-based content analysis to process a huge dataset for international and interdisciplinary comparisons. The Structural Topic Model analysis of over 7,000 abstracts reveals that domestic ELT associations (i.e. CELES, JASELE and KATE) in Japan tend to focus more on language analysis (e.g. grammar, cognitive mechanisms) and less on socially contextualised frameworks (e.g. language learner identity) than international ELT associations. In addition, researchers in the domestic ELT associations tend to employ frameworks that assume the Japanese education system but pay less attention to phenomena outside of this system. Based on these findings, this paper discusses the future directions of analysing academic trends in ELT.