探究自我調整學習教學對電腦自動批改系統中文寫作的影響

Exploring the Impact of Self-regulated Instruction (SRI) on the CSL Writing of the Automated Essay Scoring System

自我調整學習理論認為學生能夠設定自我學習目標，並可以透過有效控制自己的認知和後設認知過程， 提升學習成效（Zimmerman 1989）。此理論應用在相關學術研究的結果皆顯示，教師的介入（例如回饋策略培訓）和電腦輔助資源有助於學習者調整其學習模式，並且能夠提升目標語的學習成效。然而目前的研究仍以學習英語為大宗；相對的，針對以中文為第二語言進行寫作的研究相當少。因此，本研究透過自動作文評分系統（AES）──SmartWriting-Mandarin，探討自我調整教學模式（SRI）和非自我調整教學模式（NSRI）對中文作為第二語言（CSL）寫作的影響。具體而言，本研究使用該平臺，結合前述兩種教學模式來檢視CSL 學習者的寫作表現，並進一步調查學習者對此課程的看法。實驗課程的受試者為主修中文的越南籍學生。這些受試者皆為自願參與本教學實驗，完成SRI 和NSRI 兩種教學模式共六項寫作任務，並填寫課程調查問卷。兩種教學模式的作業皆分別以SmartWriting-Mandarin 平臺和人工的方式評分，以確認這兩種課程下的作業表現是否有顯著性的差異。研究結果顯示，SRI 和NSRI 作業之間的整體得分以及分項得分皆存在顯著性的差異，證實學習者可以透過SRI 獲得寫作技巧。此外，參與者對於透過平臺接受SRI 訓練持有積極與正面的看法。本研究結果具有寶貴的教學啟示。

Self-regulation posits that students are capable of setting up a goal and facilitating their own learning by engaging in cognitive and metacognitive processes to achieve successful outcomes (Zimmerman 1989). In light of the theory across various relevant research, instruction (e.g., feedback strategy training) and computer-assisted resources have been attested to help learners regulate their learning and enhance their competence in the target language. Although the implementation of such resources in English-related learning contexts has been extensively documented, their capacity to help L2 learners develop their writing abilities in Chinese learning contexts remains unclear. Hence, this study aimed to explore the effect of self-regulated instruction (SRI) and nonself- regulated instruction (NSRI) on Chinese-as-a-second-language (CSL) writing through an automated essay scoring (AES): the SmartWriting-Mandarin. More specifically, the study examined the writing performance of CSL learners via the implementation of the platform combined with the two instructions, and additionally explored learners’ perceptions of the CSL class. This instructional research examined an intact class comprised of Vietnamese students who were majoring in the Chinese language. The students voluntarily participated in the study, which included the completion of six writing tasks divided into two types of instruction—SRI and NSRI—as well as the completion of a questionnaire survey. Both the SRI and NSRI assignments of the participants were graded by the SmartWriting-Mandarin platform and human ratings to investigate whether there were significant differences between the two types of assignments. The results revealed that the holistic scores between SRI and NSRI assignments differed significantly. Further, the difference in analytic scores between the two types of assignments was also significant, suggesting that writing skills could be acquired through SRI. In addition, the participants had positive opinions toward SRI via the platform. The results of this research have potentially valuable pedagogical implications.