| 摘要: | 近年來，教育機構普遍引入人工智慧技術，以提升學生的學習體驗和知識內化能力。在許多學習方法中，教中學是其中一個有效的方式，研究指出教中學(LdL)的設計可以提升學生在活動中的興趣，透過讓學生擔任教學者，從而深化對所學知識的理解與應用，加強學生的積極參與。閱讀興趣對學生來說非常重要，能夠激發學生對於學習英語的熱情，並促使他們主動參與閱讀活動，特別是在國際化的環境中，英語能力成為一個關鍵的優勢，英語閱讀是學生必須培養和提升的必要能力之一。
此外，過去有許多研究探究聊天機器人對學習及閱讀成效的影響，在聊天機器人影響閱讀情意上的研究尚為少數。本研究結合英語繪本，開發出以問答對話為主要形式的聊天機器人系統，學生需理解繪本內容，在訓練機器人的過程中向機器人提問及教導它正確答案，並在同儕之間加入機器人競爭機制及排行榜的呈現，學生們互相挑戰對方所訓練的機器人，檢視他人的訓練成果，對方也能看到受挑戰的提問內容，達到提升學生參與度、互動性、即時回饋的效果。本研究實驗對象為臺灣桃園市某國民小學的95名五年級學生，實驗組的47名學生在閱讀時使用問答機器人，控制組的48名學生閱讀時不使用問答機器人，實驗期間學生每週進行一次英語閱讀活動，並設計最後一週為實驗組的決賽活動，探討學生在英語閱讀活動中若有問答機器人的加入對心流以及情境興趣是否有影響、分析學生與機器人互動時的提問情形，以及利用相關性分析來觀察學生在心流、情境興趣的情意層面中，與提問情形的關係。
研究結果顯示，在閱讀活動結束後，實驗組學生的心流隨著閱讀活動的參與次數顯著遞增，其對閱讀的專注力、好奇心、內在興趣及整體心流都顯著高於控制組學生，對閱讀的控制力上差異不大。另外在情境興趣上，顯示問答機器人能夠觸發實驗組學生的情境興趣，且不論是在維持感受、維持價值、觸發及整體情境興趣的向度上，實驗組學生均顯著高於控制組學生。相關性分析顯示學生對閱讀的情境興趣越高，其提出故事人物感受的問題數量也越多，其他的問題類型對投入狀況與情境興趣大多沒有關聯，此外發現英語能力高的學生著重於提出人物、環境及結果類型問題，英語能力低的學生更偏好探究因果關係的問題。;In recent years, many educational institutions have adopted artificial intelligence technology to enhance students′ learning experiences and promote the internalization of knowledge. Among various learning methods, Learning by Teaching(LdL) has been identified as an effective approach that can increase students′ interest and deepen their understanding and application of the learned content. Developing a strong interest in reading is crucial for students as it stimulates their enthusiasm for learning English and encourages active participation in reading activities. Particularly in the context of internationalization, English proficiency has become a key advantage, and the ability to read in English is considered a necessary skill that students must cultivate and improve.
Furthermore, previous research has explored the impact of chatbots on learning and reading outcomes, although studies focusing on the influence of chatbots on emotional engagement during reading remain limited. This study combines English picture books with a chatbot system, utilizing a question-and-answer dialogue as the primary form of interaction. Students are required to comprehend the book′s content during the picture book reading activity, ask questions to the chatbot, and teach it the correct answers. Additionally, a competitive mechanism and leaderboard are introduced among peers, allowing students to challenge each other′s trained chatbots and evaluate the performance of their peers′ training outcomes. This competition mechanism aims to enhance student engagement, interactivity, and provide real-time feedback.
The experimental study involved 95 fifth-grade students from a elementary school in Taiwan. The experimental group of 47 students used the QA Chatbots during their English reading activities, while the control group of 48 students did not utilize the QA chatbot. Throughout the experiment period, students engaged in weekly English reading activities, and arranged competition activities for the experimental group in the last week. The study aimed to investigate the impact of incorporating chatbots into English reading activities on flow experience, situational interest, student-questioning patterns during interactions with the chatbot, and the relationship between flow, situational interest, and questioning patterns using correlational analysis.
The research findings indicate that after participating in the reading activities, students in the experimental group demonstrated a significant increase in flow experience with the frequency of their engagement. They exhibited higher levels of concentration, curiosity, intrinsic interest, and overall flow experience compared to the control group, while the differences in control over reading were not significant. Furthermore, regarding situational interest, the results revealed that the QA Chatbots triggered higher levels of situational interest among students in the experimental group. In dimensions such as Maintained-SI-Feeling, Maintained-SI-Value, Triggered-SI, and overall situational interest, the experimental group outperformed the control group significantly. Correlational analysis demonstrated that students with higher situational interest generated more questions related to the characters′ “Feeling”, while other types of questions showed little or no correlation with engagement and situational interest. Additionally, it was observed that students with higher English proficiency focused more on asking questions related to “Character”, “Setting”, and “Outcome resolution”, whereas students with lower English proficiency preferred exploring causal relationships in their questions. |
