Designing CAST: A Computer-Assisted Shadowing Trainer for Self-Regulated Foreign Language Listening Practice

Document Title

Designing CAST: A Computer-Assisted Shadowing Trainer for Self-Regulated Foreign Language Listening Practice

Document Information

• Research Area: Language Learning Technology, Computer-Assisted Education
• Keywords: Computer-Assisted Language Learning, Self-Regulated Learning, Multimedia Learning, Language Shadowing Training, Audio and Speech Interfaces, Visualization Tools, Reflective Practice

Research Background and Problem

• Research Problem:

  • Shadowing is a widely popular and effective language learning technique, but existing software tools often focus on speaking skills while neglecting shadowing exercises that emphasize listening.
  • In self-regulated learning environments lacking external feedback, learners often struggle to focus on listening and effectively reflect on their learning process.

• Research Importance:

  • Listening skills are a fundamental component of language learning. Existing studies show that shadowing can effectively improve learners' listening abilities, but there is a lack of self-regulated shadowing tools specifically designed for listening development.

• Motivation and Related Work:

  • Through literature review and empirical research, the authors point out that existing shadowing systems focus more on speaking rather than listening (e.g., the "WithYou" system).
  • The authors identified through needs analysis that language learners face high cognitive load, text dependency issues, and lack effective support for reflecting on errors during self-regulated shadowing practice.

Solution

• Proposed Solution:

  • Introducing the "Computer-Assisted Shadowing Trainer" (CAST), a system specifically designed for shadowing exercises that emphasize listening.
  • Proposing four innovative design elements to support self-regulated shadowing learning:
    1. Real-Time Highlighting: Used to mark and visualize parts of the text that learners find difficult or easy.
    2. Contextual Text Blurring: Blurring parts of the text to reduce learners' dependency on it and promote self-reflection.
    3. Listening Comparator: Interactive display comparing learners' recordings with the target text for post-practice self-evaluation.
    4. Adjustable Pause Handles: Adding short pauses between sentences to reduce cognitive load while maintaining the natural rhythm of the target speech.

• Implementation Steps:

  • Designing a two-phase model based on "Listen-Practice-Reflect" and "Shadow-Practice-Reflect" to support learners in dynamically adjusting content (e.g., blur/unblur).
  • Providing mapping tools (text highlighting), pause handles, and visualization of recordings to help learners integrate reflection with practice more effectively.

Research Outcomes

• Specific Outcomes:

  • CAST enhanced learners' ability to focus on listening and effectively supported self-regulated learning behaviors.
  • Learners using CAST were able to systematically track, evaluate, and improve their shadowing performance, especially in challenging listening segments.

• Advantages Over Existing Solutions:

  • Compared to traditional tools, CAST offers a more structured system design focused on listening practice.
  • It introduces specific design elements (e.g., text blurring) that effectively address learners' text dependency and listening interference issues.

• Experimental or Evaluation Results:

  • At every stage, participants using CAST outperformed those using baseline interfaces, significantly improving listening focus, reflective abilities, and self-assessment effectiveness.
  • Participants (N=12) provided positive feedback on CAST, noting that its interactive design better supports deep engagement in learning.

• Limitations and Future Directions:

  • Limitations:

    • The current study provides limited validation of long-term learning gains, relying only on single-session and short-term evaluations.
    • Independent validation of component-level effects is lacking, such as whether individual design elements impact learning outcomes.

  • Future Directions:

    • Expanding the system to support speaking skills training by adapting the "Listening Comparator" into a "Speaking Comparator" with adjusted comparison criteria.
    • Conducting multiple long-term experiments to verify the impact of CAST on learners' long-term listening and language skill development.
    • Exploring how CAST's design principles can be applied to other language learning or multimedia learning scenarios, such as solving math problems or reciting epic poetry.

Conclusion

CAST is the first self-regulated shadowing system designed based on learners' needs, addressing text dependency issues in shadowing exercises focused on listening while promoting self-reflection and evaluation. Experiments confirm CAST's effectiveness in supporting language learners' listening focus and improving their learning experience.