探討在弱結構問題解決情境中，工作記憶廣度、後設認知能力和問題解決表現之關聯

Abstract

培養問題解決能力是科學教育中一個重要的基本能力，過去研究指出，學習者的問題解決表現與問題任務難易程度、學習者的後設認知行為以及其工作記憶廣度有關(Veenman,Prins,＆Elshout,2002；Shin,Jonassen,＆McGee,2003；Cho, Holyoak ,& Cannon, 2007)，然而有關學習者的後設認知行為表現是否會受到工作記憶廣度和問題解決任務難易度的影響，過去研究並未有較深入的探討，且任務多以類比推理為主，鮮少以問題解決為探討之情境。因此本研究目的有二：首先欲了解不同工作記憶廣度學習者之問題解決表現和後設認知行為是否受到問題難易程度影響，其次則探討後設認知能力在問題解決過程中所扮演的角色。 本研究採質量混合設計，以影響最大靜摩擦力的因素為主要概念，設計低、中、高三種不同難度之問題解決情境。研究對象為常態分班的國中二年級五個班級，全體學生先進行摩擦力概念測驗，篩選37位先備概念相當之學習者進行工作記憶廣度測驗及問題解決任務訪談，再將訪談資料進行後設認知能力編碼及問題解決能力評分，以分析問題解決表現是否與工作記憶廣度高、低有關?並探討高、低工作記憶廣度學習者於低、中、高難度之問題解決情境下，學習者的後設認知行為與問題解決表現之關聯，並於高、低工作記憶廣度組別中各選取2名個案，進行質性分析以了解在低、中、高三種不同難度下，學習者如何展現設認知行為進行問題解決。 結果顯示，學習者之問題解決表現受到問題難易度之影響，且於中難度的問題解決情境下，問題解決表現最佳，其中「分析題目的圖片和條件」及「關係圖的判斷」此兩個低層次之問題解決表現，隨難度增加並沒有太大變化；而「變因影響的關係判斷」及「統整及推論」兩個較高層次之問題解決表現於中難度任務下表現最佳。此外低、中、高三種不同難度問題解決情境下，問題解決表現皆為高廣度學習者優於中廣度學習者，中廣度學習者優於低廣度學習者，顯示學習者的工作記憶廣度大小的確會影響問題解決能力。質性的問題解決過程訪談分析則發現，高廣度學習者在面對難度較高之問題解決情境時，能監控推理過程以降低過程中的推理失誤，或於失誤時能及時覺察而回頭檢視並提出修正，顯示工作記憶廣度高之學習者有足夠的認知資源進行監控並調整學習策略進行問題解決。反之，低廣度學習者，僅於中難度之問題解決情境時，能監控推理過程以降低推理失誤發生，當問題難度提升至高難度時，則因認知負荷使過高而無法監控推理過程，甚至發生推理失誤而不自覺，降低問題解決成效。 本研究顯示，問題難易程度及後設認知行為對於問題解決表現有重要的影響。

Cultivate problem-solving ability is an important goal in science education. Previous studies indicated effects of task difficulty, learners’ metacognitive behaviors, and working memory span on problem-solving performance (Veenman, Prins, & Elshout, 2002; Shin, Jonassen, & McGee, 2003; Cho, & Cannon, 2007). Whetehr learenrs’ metacognitive behaviors are affected by their working memory span and the task difficulty remains unexplored. To address this gap, the present study has two aims, to understand whether students with different working memory span have different problem-solving performance and metacognitive behaviors, and how problem-solving performance and metacognitive behaviors vary accprding to task difficulty. This study also explore the role of metacognitive ability in problem- solving . The study took a mixed-method approach. Problems with low, medium and high levels of difficulties were designed on the topic of friction. Participants were eighth-graders who have received related instructions. A test friction was implemented to select, 37 students who have similar base of prior knowledge on Friction. These students were interviewed while completing the problemsolving tasks. Their task performances were scored and the interviews were coded for cognitive and metacognitive habavoirs. Relationships between learners’ working memory span, metacognitive behaviors, and problems-solving performance in different levels of difficulties were then analyzed. Cases were also used , to understand learners’ metacognitive behaviors and problem-solving processs in different task difficulties. The results show that learner's’ problem-solving performance varied depended on task difficulty. Performance on “analysis of task conditions”and “chart decomposion”did not vary across task diffulcities; however, learners exhibited optimal performance at the medium difficult level for performance on “reasoning causal relations”and “integrating and inferring”. tudents with large working memory span outperform their counter cohors on all diffiultiy levels. Qualitative findings also showed that learners with higher working memory span were abled to monitor problem-solving process to reduce reasoning flaws and/or made corrections with errors occurred, which in turn, yielded better performances. Learenrs with lower working memory span demonstrated the aforementioned behavioral pattern only on the task with midiun level of difficulty.problem-solving perofmrance was droped at the high difficulty level due to cognitive overload. In that situation, learenrs no longer monitored their reasoning process and were not aware of realsoing flaws. This study shows that task difficulty and emtacognitive behaviors have great influence on problem- solving process and performance.