The effects of automatic speech recognition (ASR)-based peer feedback on EFL diphthong pronunciation improvement and retention

Abstract

The present study examined the effects of automatic speech recognition (ASR)-based peer feedback on Egyptian English-majored students’ EFL diphthong pronunciation improvement and retention. The participants comprised 28 first-year students in the English Language Department, Faculty of Archaeology and Languages, Matrouh University, throughout the Spring semester of the academic year 2023/2024. They were randomly assigned into a control group (n = 14) which received traditional pronunciation instruction and an experimental group (n = 14) which used ASR-based peer feedback. An EFL diphthong pronunciation test was devised and administered as a pre-, post-, and delayed posttest. Semi-structured interviews were also held to ascertain the students’ perceptions toward the use of ASR-based peer feedback in their diphthong pronunciation training. For three weeks, students completed the assigned in-class and out-of-class activities where they practiced producing the English diphthongs using the ASR application Speechnotes with peer feedback. Results indicated that the experimental group significantly surpassed the control group on the posttest and the delayed posttest. Thus, ASR-based peer feedback had significant effects on students’ diphthong pronunciation improvement and retention. The students’ responses also showed positive perceptions toward its use. Such results suggest that Egyptian EFL students could improve their diphthong pronunciation through using ASR-based peer feedback.