Scholarlink

<p>This mixed-methods study rigorously evaluates the efficacy of artificial intelligence (AI)-enhanced vocabulary learning strategies—incidental exposure, contextual priming, and multimodal scaffolding—relative to traditional instructional approaches in English as a Foreign Language (EFL) pedagogy. Employing a pretest–posttest randomized controlled trial with 383 Chinese EFL learners, the research quantifies AI’s impact on both immediate lexical acquisition and long-term retention, while qualitatively elucidating learner perceptions of engagement and instructional value. Quantitative findings indicate that AI-mediated multimodal strategies yield significantly higher vocabulary gains (posttest <i>M</i> = 137.00, <i>SD</i> = 5.51) and sustained retention (delayed posttest <i>M</i> = 129.00) compared to contextual (<i>M</i> = 121.00), incidental (<i>M</i> = 113.00), and control groups (<i>M</i> = 76.05), with robust multivariate effects (Pillai’s Trace = 0.766, <i>p</i> &lt; .001, partial η² = 0.589). Thematic analysis of semi-structured interviews highlights AI’s strengths in delivering personalized feedback, adaptive content, and immersive contextualization (e.g., domain-specific VR simulations), while also surfacing potential drawbacks such as sensory overload (β = –0.53) and cultural bias in AI-generated materials. Interpretation of these results is explicitly anchored in four complementary educational and pedagogical frameworks: (1) Sociocultural Theory, which positions AI as a mediating artifact within the learner’s Zone of Proximal Development; (2) the Cognitive Theory of Multimedia Learning, which accounts for the superior efficacy of multimodal approaches via dual-channel (visual-auditory) processing; (3) the Involvement Load Hypothesis, which predicts vocabulary gains based on the cognitive depth and motivational salience of AI-mediated tasks; and (4) Nonlinear Dynamic Language Learning Theory, which conceptualizes vocabulary development as an emergent, context-sensitive process shaped by dynamic learner–technology–task interactions. These theoretical underpinnings collectively elucidate how AI-driven strategies optimize encoding, retention, and learner agency, while also explaining observed phenomena such as motivational decline following mastery in rigid adaptive systems (HR = 1.8). The study advances robust empirical evidence for AI’s transformative potential in vocabulary pedagogy, demonstrating its capacity to bridge retention gaps and foster learner-centered, contextually adaptive instruction. Practical implications advocate for the design and deployment of ethically transparent AI tools that balance the affordances of multimodal engagement with the necessity of cognitive load regulation. Theoretical contributions further refine nonlinear, integrative models of lexical acquisition in technology-mediated environments, offering a nuanced framework for future research and practice.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing EFL vocabulary acquisition: a randomized controlled mixed-methods investigation of artificial intelligence-driven incidental, contextual, and multimodal strategies

  • 期刊论文

  • 发表日期：

    2025年10月27日

查看全文
  • Miao Yu

摘要

This mixed-methods study rigorously evaluates the efficacy of artificial intelligence (AI)-enhanced vocabulary learning strategies—incidental exposure, contextual priming, and multimodal scaffolding—relative to traditional instructional approaches in English as a Foreign Language (EFL) pedagogy. Employing a pretest–posttest randomized controlled trial with 383 Chinese EFL learners, the research quantifies AI’s impact on both immediate lexical acquisition and long-term retention, while qualitatively elucidating learner perceptions of engagement and instructional value. Quantitative findings indicate that AI-mediated multimodal strategies yield significantly higher vocabulary gains (posttest M = 137.00, SD = 5.51) and sustained retention (delayed posttest M = 129.00) compared to contextual (M = 121.00), incidental (M = 113.00), and control groups (M = 76.05), with robust multivariate effects (Pillai’s Trace = 0.766, p < .001, partial η² = 0.589). Thematic analysis of semi-structured interviews highlights AI’s strengths in delivering personalized feedback, adaptive content, and immersive contextualization (e.g., domain-specific VR simulations), while also surfacing potential drawbacks such as sensory overload (β = –0.53) and cultural bias in AI-generated materials. Interpretation of these results is explicitly anchored in four complementary educational and pedagogical frameworks: (1) Sociocultural Theory, which positions AI as a mediating artifact within the learner’s Zone of Proximal Development; (2) the Cognitive Theory of Multimedia Learning, which accounts for the superior efficacy of multimodal approaches via dual-channel (visual-auditory) processing; (3) the Involvement Load Hypothesis, which predicts vocabulary gains based on the cognitive depth and motivational salience of AI-mediated tasks; and (4) Nonlinear Dynamic Language Learning Theory, which conceptualizes vocabulary development as an emergent, context-sensitive process shaped by dynamic learner–technology–task interactions. These theoretical underpinnings collectively elucidate how AI-driven strategies optimize encoding, retention, and learner agency, while also explaining observed phenomena such as motivational decline following mastery in rigid adaptive systems (HR = 1.8). The study advances robust empirical evidence for AI’s transformative potential in vocabulary pedagogy, demonstrating its capacity to bridge retention gaps and foster learner-centered, contextually adaptive instruction. Practical implications advocate for the design and deployment of ethically transparent AI tools that balance the affordances of multimodal engagement with the necessity of cognitive load regulation. Theoretical contributions further refine nonlinear, integrative models of lexical acquisition in technology-mediated environments, offering a nuanced framework for future research and practice.