Academic Journal
The impact of AI-enhanced interactive learning environments on EFL teachers' computational thinking pedagogical knowledge: a multi-arm longitudinal experimental study.
| Τίτλος: | The impact of AI-enhanced interactive learning environments on EFL teachers' computational thinking pedagogical knowledge: a multi-arm longitudinal experimental study. |
|---|---|
| Συγγραφείς: | Xu, Xiang1 (AUTHOR) Xxwangyi_6626@126.com, Bahari, Akbar2 (AUTHOR) |
| Πηγή: | Interactive Learning Environments. Oct2026, Vol. 34 Issue 8, p5481-5508. 28p. |
| Θεματικοί όροι: | *Computational thinking, *Language teachers, *Design, *Instructional systems, *Cultural competence, *Pedagogical content knowledge, *Longitudinal method |
| Περίληψη: | Despite growing AI integration into language education, evidence on how AI-enhanced interactive learning environments (AI-ILEs) cultivate EFL teachers' computational thinking pedagogical knowledge (CTPK) remains limited. This explanatory sequential mixed-methods study embedded a multi-arm randomized controlled trial within five Chinese universities. Quantitatively, 125 postgraduate EFL teachers were randomly assigned (1:1:1:1:1; block randomization stratified by site, experience, gender) to four AI-ILE interventions—Instructional Decomposition/Collaborative Planning, Pattern Recognition/Formative Assessment, Conceptual Abstraction/Mapping, Algorithmic Sequencing/Debugging—or non-AI control. CTPK was measured at three timepoints. MANCOVA with multiple imputation and inverse-probability weighting addressed missing data and attrition. Qualitatively, semi-structured interviews (n = 20; maximum-variation sampling) underwent dual-cycle thematic coding, supported by intercoder agreement and member checking. Integration occurred through joint display matrices generating meta-inferences. Results indicated clear intervention effects, with Decomposition/Collaborative Planning demonstrating the largest sustained CTPK gains. Qualitative themes revealed enhanced diagnostic precision requiring cultural calibration, professional identity shifts necessitating institutional support, and sustainability linked to collaborative communities. Meta-inferences showed convergence between quantitative efficacy and qualitative confidence gains, with divergence highlighting that efficacy alone does not ensure implementation without addressing algorithmic bias. Findings support prioritizing decomposition-based AI-ILEs while mandating culturally responsive design for scalable, equity-centered implementation. [ABSTRACT FROM AUTHOR] |
| Βάση Δεδομένων: | Academic Search Index |
καταχωρήστε σχόλιο πρώτοι!