科學文本閱讀之理解監測準確度對自我調節學習策略及閱讀理解表現的影響

Abstract

科學文本的閱讀是學習科學概念的一個重要且基本的方式，學習者藉由閱讀和理解科學文本的內容，建立心智模型，並應用至新的情境。而在複雜的學習任務中 (如科學閱讀或問題解決)，自我調整學習 (self-regulated learning, SRL) 策略的使用，對於學習者在學習知識的過程相當的重要。過去的研究雖已分別指出理解監測準確度或自我調節學習策略使用的類型和品質會影響學習者的閱讀理解表現，但卻少有研究探討此三個因素之間的關係，也未對學習者對閱讀的理解監測準確度如何影響學習者自我調節學習策略的選擇，影響閱讀理解表現。故本研究之目的在於分析理解監測準確度如何影響學習者自我調節學習策略的使用，進而影響其閱讀理解表現。 本研究採質量混合設計，探討非生物主修之大學生於網路平台進行科學文本線上閱讀的自我調節學習行為和表現。本研究以學習者在學習歷程中展現的自我調節學習策略之類型和頻率、理解監測準確度和閱讀理解表現進行相關分析，進一步將學習者以理解監測準確度之表現分為高、中、低三組，比較三組之學習者在閱讀理解表現的差異，以及比較三組學習者於跨段落和段落內閱讀文本時，自我調節學習策略使用頻率與運用方式的差異，最後對此三組學習者中各選取兩名進行質性分析，以了解其閱讀歷程中如何展現其自我調節學習策略，進而影響閱讀理解之表現? 理解監測準確度、閱讀理解表現與自我調節學習策略使用頻率的相關性分析結果顯示，理解監測準確度與閱讀理解表現呈顯著正相關，且學習者在認知策略中，總結、推論和重新閱讀等認知策略的使用頻率與閱讀理解表現直接相關，其使用頻率越高，學習者的閱讀理解表現越好。而後設認知策略中僅有評價所讀內容等後設認知策略的使用頻率與閱讀理解表現直接相關，其使用頻率越高，學習者的閱讀理解表現越好。 理解監測準確度高、中、低組於閱讀理解表現及自我調節學習策略使用頻率的分析結果顯示，理解監測準確度越高的組別，其閱讀理解表現越好，使用提出推論策略的頻率越高。進一步比較三組學習者於段落內和跨段落的自我調節學習策略，發現三組學習者在理解段落內訊息時，無論是認知策略或是後設認知策略，使用的頻率差異不大，顯示這三組的學習者都需要先於段落內使用自我調節學習策略以達到文意理解；三組學習者的主要差異顯現在跨段落的策略使用，理解監測準確度較高的學習者，在自我調節學習策略使用上，除會在段落內使用外，亦較常跨段落使用提出推論及提出總結等認知策略，並會使用跨段落的覺察閱讀理解、進行學習判斷等後設認知策略來評估其理解內容和成果；而理解監測準確度低的學習者則較少在跨段落的閱讀展現提出推論及提出總結等認知策略，以及運用覺察閱讀理解、進行學習判斷等後設認知策略。質性資料分析結果則支持分組比較的結果，且進一步顯示理解監測準確度高的學習者，在進行閱讀時，相較於理解監測準確度低的學習者，會主動監測自己的學習，透過監測蒐集與學習相關的資訊，對照自己對任務的標準，決定是否啟動自我調節學習機制調整認知策略，以促進學習者形成較完整的心智模式。理解監測準確度低的學習者則因無法有效運用後設認知策略檢核其理解表現和認知策略使用之效益，雖能達文意理解和理解段落內的概念，但較難形成跨段落之概念整合，無法進一步建立概念間的推論，心智模式較不完整。 研究結果可知，不僅是認知和後設認知策略的使用品質和類型會影響閱讀理解表現，提出推論、提出總結等認知策略和評價所讀內容、覺察閱讀理解及進行學習判斷等後設認知策略，學習者在閱讀歷程中所進行的推論、監測的尺度 (使用於段落內或是跨段落) 和時機，也會影響學習者在閱讀表現上僅達文意理解或是能夠提升至情境理解的層次。

Reading a scientific text is an major and important way of learning science concepts. Through reading, students comprehend the content of the scientific text, construct a mental models, and may apply it to a new situation. Self-regulated learning (SRL) is essential when learning through or solving complex tasks (e.g., science reading or problem-solving). Extant literature has shown that, although separately, calibration accuracy or quality of SRL strategies learners use would affect reading comprehension; however, the interrelationships among the three variables were unexplored. To address this gap, the aim of the study was to investigate how calibration accuracy may affect learners’ use of SRL strategies, which in turn influences their reading comprehension. A mix-method approach was used to explore non-biology majored college students’ calibration accuracy, SRL strategies, and reading comprehension when reading a biology text in a computer-based learning environment. Correlations among calibration accuracy, reading performance, and use of SRL strategies were first explored. Reading comprehension and frequencies of SRL strategies used within and across text paragraphs were then compared for students with high, moderate, or low calibration accuracy. Two case students from each calibration accuracy group were used to exemplify how SRL strategies unfolded during science reading would influence reading comprehension. The findings of correlation analyses revealed a significant positive correlation between calibration accuracy and reading comprehension. Students who deployed higher frequencies of “summarization”, “inference”, “re-reading” and “content evaluation”, exhibited better reading comprehension. Between-group comparisons indicated that no statistically significant differences were found on students’ use of cognitive and metacognitive strategies when reading within paragraphs. However, significant differences were found for inter-paragraph SRL strategies. Learners with better calibration accuracy used SRL strategies to comprehend sentences and paragraphs; they also utilized cognitive strategies such as “inference” and “summarization” to infer across paragraphs. At the meantime, they used metacognitive strategies such as using “feeling of knowing” and “judgment of learning” to evaluate inferences resulted from across-paragraph reasoning. On contrary, learners with poor calibration accuracy mainly used “inference”, “summarization”, “feeling of knowing”, and “judgment of learning” for reading within paragraphs rather than across paragraphs. Qualitative analyses for the case students supported the results of group comparisons and illustrated characteristics for dynamic uses of SRL strategies when reading within and across paragraphs. The findings of the study indicated that reading comprehension is influenced by frequencies of particular cognitive and metacognitive strategies; the timing and reading scale (within or across paragraphs) in which the students’ SRL strategies were activated would also affect their level of reading comprehension.