Quantifying the Benefits of the Reread- Before-Answer Strategy in Japanese High- School EFL

Abstract

Digital e-book platforms collect detailed logs of students' reading behaviors, yet it remains unclear whether brief revisits of previously viewed pages before answering open-book quiz questions significantly improve outcomes. This study investigates whether "Reread-Before-Answer" (RBA) loops, brief backward navigation immediately preceding quiz responses, impact quiz performance. Analyzing one month of BookRoll logs from 263 Japanese high-school English learners, we applied sequence mining to detect RBA loops and logistic regression to quantify their effects, controlling for session duration, navigation length, and prior English proficiency. Results show RBA loops occurred in approximately 40% of sessions and consistently predicted better performance, with learners who engaged in these loops showing significantly higher quiz accuracy compared to strictly linear readers. Pattern analysis further revealed that successful students preferred concise review behaviors over prolonged rereading. These findings highlight that encouraging short, intentional revisits to relevant content just prior to assessment can effectively enhance students’ quiz performance. Consequently, integrating structured rereading prompts within digital textbooks could serve as a practical intervention for educators and instructional designers to optimize learning outcomes.