シャドーイングにおける音声認識フィードバックがスピーキング流暢さに与える影響

The author reports on an experiment using speech-to-text (STT) technology in gamified shadowing activities. Pre- and post-treatment speaking characteristics of Japanese EFL learners were analyzed. Learners engaged in short- to medium-length in-app shadowing sessions on mobile phones, while control groups completed non-digital activities of similar duration. Data on pauses, syllables, utterance time, phonation time, and speech rate were used to calculate speed and breakdown fluency. STT provided customized feedback after each attempt, encouraging error correction and repetition, with immediate and cumulative progress tracking. Part one of the study quantified fluency patterns in low- and intermediate-level speakers; part two compared STT treatment effects at 5 and 20 hours against two control groups. The findings contribute to understanding how mobile features like STT and voice recognition affect EFL learners’ speaking fluency.