Error gravity is an important feature that has recently gained attention in writing research, based on suggestions that it contributes to the calculation of accuracy metrics in writing performance and the selection of instructional methods. However, the findings from error gravity studies have been inconsistent. Our systematic review critically examined the research methodologies employed in error gravity studies, focusing on three key aspects: task conditions, error categories, and assessment criteria. Of the 92 studies retrieved from two databases, 17 studies were selected after excluding duplicates and studies that did not align with the purpose of this review. Six studies were further analyzed following a coding process. The findings revealed that writing studies utilizing error gravity can be broadly classified into two types: one focusing on one type of error in a sentence, and the other aiming to calculate the score of a written accuracy measure. Furthermore, a detailed analysis indicated inconsistencies in the three methodological settings, making it difficult to compare and synthesize the research results. This study also proposed improvements for each of the three key aspects to further advance error gravity research.

抄録全体を表示

This study explored the potential of automatic speech recognition (ASR) transcription for linguistic research using a preliminary analysis of transcripts generated by three ASR systems: Google Cloud Speech-to-Text, Rev AI, and Whisper. The analysis included 214 sample files of spontaneous speech from Japanese learners in the International Corpus Network of Asian Learners of English corpus. OpenAI was employed to remove disfluencies from the transcripts, which resulted in highly accurate pruned versions. The performance of the ASR systems was evaluated using these pruned transcripts, with word error rate (WER) serving as the evaluation metric. Among these systems, Whisper exhibited the best performance, achieving a WER of 28.7% for pruned transcripts and 18.8% for pruned-normalized transcripts. This study also identified challenges for ASR in transcribing spontaneous non-native speech. The three systems, including Whisper, exhibited higher WERs for non-native speakers than for native English speakers. In addition, all three systems tended to correct erroneous morphological forms in verbs and nouns, with Whisper showing the most prominent trend, suggesting that ASR transcripts generated by the current systems, particularly Whisper, may not be suitable for error analyses involving morphological forms.

抄録全体を表示

Game-based Classroom Response Systems (GCRS) such as Kahoot! have attracted attention for promoting interactive learning. Although Kahoot! is widely used, recent studies suggest that its effectiveness for second language (L2) vocabulary acquisition may be limited (e.g., Reynolds & Taylor, 2020). This is potentially due to suboptimal utilization for vocabulary learning. To address this issue, the present study examined the differential effects of two game modes on the effectiveness of vocabulary learning. Sixty-seven junior high school students were assigned to either the “Treasure Trove Mode” group, which offers repeated exposure to target items, or the default “Classic Mode” group. Results from a Generalized Linear Mixed Model (GLMM) showed that the Treasure Trove Mode group significantly outperformed the Classic Mode group in an immediate post-test, aligning with previous research that highlights the benefits of within-session repeated retrieval (e.g., Nakata, 2017; Peters, 2014). However, no significant differences were found between the groups in a delayed post-test one week later, suggesting that a single session of repeated retrieval alone may not ensure long-term retention. These findings underscore the short-term effectiveness of repeated retrieval in Kahoot!’s Treasure Trove Mode, while also indicating the need for additional strategies, such as increased repetition frequency or distributed practice, to achieve lasting vocabulary retention.

抄録全体を表示

This pedagogical article explores the integration of generative AI and audio journal homework as a supplementary activity for a university English as a foreign language (EFL) speaking course. The method involves writing a journal in the students’ first language, AI-assisted translation, self-review of the translated script, recording and submitting an audio file, and participating in structured in-class reporting sessions to help address difficulties in expressing daily experiences and ideas in English. By providing students with access to conversational expressions through the support of generative AI and opportunities to practice spoken English within a scaffolded framework, the method indicated potential in enhancing practical English use. End-of-course surveys highlighted positive feedback on students’ perceived achievements, especially in practical English use, while progress in pronunciation and anxiety reduction was less pronounced. I observed that students exhibited proactive involvement in refining AI-assisted translations, which deepened their understanding of linguistic nuances. While the method shows potential for bridging gaps in conversational English, additional scaffolding—such as pronunciation guidance and low-pressure conversational practice—may be necessary to support improvements in pronunciation and accent reduction. Such refinements are expected to make the method a valuable addition to formal language studies.

抄録全体を表示