Differential effectiveness of text-to-speech technology on secondary school students’ L2 speaking performance: A mixed-methods study

Abstract

This study explored the impact of text-to-speech (TTS) technology on the English-speaking skills of secondary school students across different proficiency levels and investigated the factors influencing its effectiveness. The research employed a mixed-methods approach, combining pre- and post-test speaking assessments with semi-structured interviews and focus groups. Quantitative results revealed a pattern of differential effectiveness: the quasi-experimental group demonstrated significant gains overall when compared to their baseline, and between-group comparisons showed that this improvement was significantly greater than that of the control group across all four cumulative ability groupings. The effect was most pronounced among low-achieving students, while mid- and high-achieving students showed more modest gains. Qualitative findings elucidate this differential pattern, indicating that low-achieving students, who faced significant comprehension barriers and scarce learning resources, perceived TTS as a critical, compensatory tool. In contrast, high-achieving students found TTS misaligned with their goals for authentic communication and nuanced feedback. Although all students expressed a preference for human teachers, perceived usefulness was the primary driver of engagement. The findings suggest that, within the context of prepared read-aloud tasks, TTS technology may offer targeted support that could contribute to more equitable learning outcomes for low-achieving learners, though its effectiveness is mediated by learner proficiency, perceived task difficulty, and the availability of alternative resources. The results provide initial evidence that successful implementation likely requires pedagogically differentiated strategies that align the tool's functionality with specific learner needs.