與他人共同合作解決問題，是21世紀的關鍵能力，而科學更是各國強盛的關鍵，然而科學相關問題必然涉及諸多變項之交互影響與複雜的科學概念，學習者可以透過合作調整變數進行電腦模擬，觀察過程幫提升解決問題的能力。 本研究目的在了解學習者在合作問題解決中所表現的歷程，進一步探討學習者不同合作品質與不同學習方法的學習者歷程差異，協助教師能培養學習者合作問題解決能力。 本研究受試者為高中學生共24名，用簡報教學線上合作科學問題解決系統操作，時間為一小時，對所有學習者施以一小時的「合作問題解決概念教學」，授課內容著重於引導學習者成員嘗試透過對於意義的共同建構、反思、溝通、修正來解決問題。後以一活動請學習者實際操作，時間亦為一小時，之後正式給予受試者一簡單難度活動與一困難活動，其活動時間各為一小時，應用電腦模擬協助學生進行合作科學問題解決活動，透過蒐集受試者在電腦內的歷程資料，探討學習者合作科學問題解決的行為模式，使用PISA2015 合作問題解決評量架構做為分析標準，透過運用滯後序列分析對話串內容，討論其行為模式之間的差異，本研究亦將學習者分別依其學習方法與在合作問題解決活動中所感受的合作品質進行分組，進一步比較不同組別學習者的合作問題解決歷程之差異。 研究結果顯示學習者經歷合作問題解決之概念教學後，在簡單的問題解決動中，其較容易形成假設。而在低合作品質小組之學習者在問題解決歷程中需要較多的溝通建立共同的理解。而深層學習方法組之學習者在問題解決歷程中著重假設進而解決問題，而淺層學習方法組之學習者透過討論套用公式與建構共同理解方式去解決問題。最後亦對未來之研究方向提供建議。 ;Knowing how to solve problems with peers collaboratively has become one of the most important skills in the 21st century. Obviously, science-related problems-solving concerns the interaction of many variables and complex scientific concepts. To enable the learners to have the ability of problems-solving, learners need to learn to adjust variables and observe the process in the context of computer simulations. The purpose of the study was to explore the learners’ differences in their behavioral patterns of problem solving and different learning methods. The results might be used to help teachers acquire the ability to help learners the ability to solve problems collaboratively. The participants of this study were 24 senior high school students who were in the context of computer simulations to solve assigned problems. First, they went through one-hour collaborative scientific problem-solving activities. The activities focused on guiding the participants to solve problems by constructing, reflecting, communicating, and modifying. Then the students used the theory they learned from the aforementioned activities to finish a one-hour practice activity. At the end, they were given a simple and a hard activity to do in an hour respectively. This study adopted the conceptual framework of Collaborative Problem solving skills for PISA 2015 to conduct data analysis through the use of Lag Sequence Analysis（LSA）of the conversation cluster. This study also categorized the student into different groups based on approaches learning, as well as their perceined collaboration quality doing Collaborative Problem solving activity. The results showed that learners were more likely to construct hypotheses after they experienced the instruction of collaborative problems-solving. In the problem-solving process, the low-quality of collaboration group tented to spend more time on communication to build a common understanding than the high-quality of collaboration group. The group of deep learning approaches focused on the hypotheses and solved the problems in the course of problems-solving, while the group of superficial learning approaches tried to solve the problems by discussing the formulas and constructing the common understanding. Finally, suggestions for the future research are provided.
