| Project/Area Number |
19K00803
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 02100:Foreign language education-related
|
|---|
| Research Institution | Teikyo University (2021-2023)
Tokai University (2019-2020) |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ボビー ヒロユキ 九州産業大学, 語学教育研究センター, 准教授 (20536247)
|
|---|
| Project Period (FY) |
2019-04-01 – 2025-03-31
|
| Project Status |
Granted (Fiscal Year 2023)
|
| Budget Amount *help |
¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2022: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2021: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2020: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2019: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
|
|---|
| Keywords | speaking fluency / elicited imitation / AI language assessment / educational technology / comprehensibility / speed fluency / breakdown fluency / shadowing / STT / MALL / speech to text / TEL |
|---|
| Outline of Research at the Start |
We will quantify the effect of short to moderate duration usage of Speech recognition technology on speaking characteristics of Japanese EFL learners. This will be carried out on mobile phones in a customized English learning app.
This research will help further our understanding of the effects of advanced mobile features such as Speech-to-Text, Voice-recognition embedded in highly interactive English learning activities by quantifying their effects. This will aid in the design of more effective mobile English learning tasks for EFL learners.
|
| Outline of Annual Research Achievements |
Our research this year has focused on integrating AI-driven analysis with traditional human evaluations to assess ESL learners' pronunciation and fluency. By using RASCH analysis on elicited imitation data, we aimed to identify differences and similarities between AI evaluators and human raters, potentially reshaping language assessment methods for enhanced consistency and scalability.
At the 2023 JALT National Conference, we highlighted the importance of comprehensibility and speaking evaluation in ESL education and demonstrated various data collection techniques for classroom application. We discussed both perception-based assessments and phonetic transcription, utilizing tools like the IPA to enhance pronunciation accuracy and integrate these methods into teaching practices effectively.
Additionally, our TnT Conference workshop focused on computer-based analysis for speaking fluency evaluation, showing how technology can offer better evaluation tools. This integration can significantly boost the efficiency of speaking training.
At the PanSIG Conference, we revisited elicited imitation techniques, which have been overshadowed by newer technologies but were preferred by students over traditional conversation practices according to our survey. This method improves pronunciation and fluency, suggesting a need to revitalize such techniques in modern ESL classrooms.
These studies collectively advocate for a blend of technology and traditional methods in ESL education, highlighting the need for adaptive teaching strategies to meet learners' evolving needs.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
This year, our research has successfully integrated AI-driven analysis with traditional human evaluations to assess ESL learners' pronunciation and fluency. Using RASCH analysis on elicited imitation data, we've identified key differences and similarities between AI evaluators and human raters, which has helped to refine and potentially reshape language assessment methods towards enhanced consistency and scalability.
We have also developed and demonstrated various data collection techniques, including perception-based assessments and phonetic transcription. These methods have been effectively integrated into teaching practices, significantly improving pronunciation accuracy. Additionally, our focus on computer-based analysis has shown how technology can provide better evaluation tools, boosting the efficiency of speaking training.
Feedback from students indicates a strong preference for our revitalized elicited imitation techniques over traditional methods, underscoring their effectiveness in improving pronunciation and fluency. Collectively, our research supports a balanced approach that merges technology with traditional methods, emphasizing the need for adaptive teaching strategies to meet the evolving needs of learners.
|
| Strategy for Future Research Activity |
Moving forward with our research project, we plan to further utilize our comprehensive data sets on fluency and speech performance, alongside our elicited imitation data, to deepen our evaluation of AI's efficacy in assessing speaking performance and fluency. This continued analysis aims to refine our understanding of how AI can complement and enhance traditional evaluation methods, ensuring more robust and scalable assessment techniques in the field of language learning.
In addition to our analytical efforts, we will actively disseminate our findings within the academic and professional communities. We intend to present our latest research outcomes at upcoming conferences. Moreover, we are in the process of drafting an academic paper that will encapsulate our recent discoveries and theoretical advancements.
Furthermore, we are committed to expanding our network within the speech evaluation community. By engaging with other researchers in this field, we hope to foster collaborative relationships that can lead to joint research initiatives. Through these strategic partnerships, we aim to not only enhance the scope and impact of our research but also to contribute to the broader goal of improving ESL pedagogy through cutting-edge technology and evidence-based practices.
|
