Technology-Based Settings for Pronunciation Skills

Meta-Analytical Insights

Abstract

This meta-analysis examines the effectiveness of technology-assisted pronunciation training (TAPT) in second language (L2) learning, with a focus on instructional settings and technological tools as moderators. Across 37 experimental and quasi-experimental studies (N = 1,924), TAPT showed a medium-to-large effect (Hedges’ g = 0.68, 95% CI [0.54, 0.82]). Classroom-based interventions (g = 0.78) produced the strongest outcomes, followed by online (g = 0.60) and laboratory contexts (g = 0.52). Mobile applications (g = 0.81) and automatic speech recognition (ASR; g = 0.73) outperformed video and multimedia resources. Although mild publication bias was detected, the high fail-safe N (785) supports the robustness of the findings. Extending earlier reviews, this study isolates contextual and technological moderators, showing TAPT is most effective when embedded in classrooms and supported by interactive, feedback-driven tools. The findings highlight the need to align pedagogy and technology: teachers can integrate TAPT into lessons, curriculum designers should prioritize ASR and mobile-based feedback, and ed-tech developers should create adaptive, interactive systems. Future research should move beyond English L2 contexts, examine learner proficiency and intervention duration, and adopt pre-registered designs to minimize bias.