Toward a Readability Index for Japanese Learners of EFL

In our previous research a linear readability formula was developed through a series of multiple regression analyses using four independent variables: (1) sentence length, (2) word length, (3) textbook-based word difficulty and (4) textbook-based idiom difficulty, and one dependent variable: year level of EFL textbook. The present study attempts to develop a new readability formula that includes the subjective judgment for each sentence of EFL texts by experienced Japanese EFL teachers who are familiar with the EFL textbooks and teaching at middle-grade schools in Japan. A multiple regression analysis was computed, using this factor as a new dependent variable and the previously obtained linear equation as an independent variable. The resultant linear equation was mathematically adjusted in order to provide a more powerful quadratic equation. The computation, which was carried out using Fukui’s College Analysis, a computer program for statistical computation, yielded a quadratic equation formula which proved to be much more powerful in predictive power (r) than the linear equation that had been derived in the previous study.