建置於網路合作學習系統中提升討論質量的自動適性介入機制

Abstract

本研究的目的是以一個網路合作學習系統作為基礎平台，設計、實作能夠提升學生討論質量的自動適性介入機制。為了設計即時判斷學生特性的系統功能，本研究利用群集分析、區別分析處理學生在網路合作學習活動中討論的資料，得到針對缺乏參與、過度協調、喜愛社交、任務工作導向等四種不同特性的分類公式，本研究將公式整合成學生模型，並依此設計判斷學生特性功能。接著，本研究參考專家教師輔助不同特性學生的專家知識，設計四個元件分別輔助不同特性的學生，各元件的功能包含尋找介入手段、決定介入手段以及將指導語客製化。為了評估介入機制，本研究安排兩組學生實際在網路合作學習平台中合作，其中一組接受介入，最後分析有介入機制、無介入機制兩組學生行為的資料。結果顯示，由於介入機制中判斷學生特性功能的準確率不佳，因此介入機制僅能輔助少部分的學生，但有超過一半的學生在問卷調查中表示對介入機制的指導語感到滿意，且有接受介入的學生較未接受介入的學生有較多知識建構類發話、較少非作業相關對話。The purpose of this study was to construct an automatic adaptive intervention mechanism that promotes fruitful discussions in a computer supported collaborative learning system. In order to identify four types of students in real-time, namely less contributing, coordination emphasising, communicative and task-oriented, the discussion data was processed by cluster analysis and discriminant analysis to derive four discriminant functions to identify the four types of students. These discriminant functions were integrated into one student model, upon which the system function to identify student types was designed. With reference to expert knowledge, this research also designed functions including intervention finding, intervention deciding and advice generation. To evaluate the intervention mechanism, two groups of students were arranged to learn within the system: one with help from the intervention mechanism, the other without. After the system log was analyzed, results indicated that the accuracy of the function to identify which type the students belong to was low. Therefore intervention mechanism can only help few students. However, most of the students indicated that they were satisfied with the intervention they received, and the students with intervention had much more discussion on knowledge construction and less on social events than the students who received no intervention.