分組合作學習大部分以學業成就高低分組，而本研究在翻轉教學模式以自行分組、隨機分組、學習風格分組等方式進行分組合作學習，目的在於探討不同分組方式對合作學習態度以及學習成效造成之差異。本研究以北部某技術學院幼兒保育系進修部二技所開設之特殊兒童心理與教學為研究實施科目，研究對象共計74位。本研究採取混和研究法，量化資料以SPSS分析學習風格與自行分組之合作學習態度及學習成效差異；質性資料則以紮根理論進行分析。本研究結果發現：1.分組方式對合作學習態度造成顯著性差異（p<.05），且自行分組高於隨機分組及學習風格分組；2.學習任務的難易度對於合作學習態度也有顯著影響（p<.05），學習任務容易者高於困難者；3.以學習風格類型分組時對合作學習並沒有顯著性的差異；4.翻轉教學模式下不同分組方式的合作學習對於學習成效都有顯著性的提升；5.學生認為不同分組方式所形成的互賴程度、溝通方式、團隊向心力、以及異質性的截長補短都不相同，故組員間的熟悉程度極為重要；6.當學習任務複雜時須先分工並建立討論群組，且同時有組內互評機制以強化合作學習態度的共學共好、責任分攤之特色。 Grouping approach in group cooperative learning is usually based on academic achievement. Rather than others, this study used self-grouping, random grouping, and learning style grouping approaches to analyze the differences in students' attitudes toward cooperative learning and learning effectiveness. In this study, 74 subjects were studied and implemented by the Institute of Advanced Education of the Department of Early Childhood Care of a technical college in northern Taiwan. This study adopted a mixed research method, and the quantitative data were used to analyze the differences in different grouping approaches toward the attitudes of cooperative learning and learning effectiveness by SPSS. The grounded theory approach analyzed the qualitative data. The results of this study found that: 1. the self-grouping was significantly higher than the random grouping and learning style grouping on the attitude of cooperative learning (p<.05); 2. the lower difficulty level of the learning task was significantly higher on the attitude of cooperative learning (p<.05); 3. learning styles did not make any significant difference on the attitude of cooperative learning; 4. Different grouping approaches with the flipped teaching model had a significant influence on the learning effectiveness; 5. Different grouping approaches had their own reliable relationship, solidarity forming, communication styles, and heterogeneity of the group members, which make the acquaintance between group members important; 6. when the learning tasks became more difficult, assigning work in advance, having a discussion network, and establishing peer evaluations in the group were suggested.
