利用眼動追蹤技術和眼動回溯自陳探討不同先備知識的大學生其評鑑科學解釋的認知歷程

培養評鑑科學解釋的能力是科學教育的主要目標之一，學習者的科學解釋評鑑表現會受到評鑑歷程的注意力分配和認知與後設認知策略使用之影響，然而過去少有研究深入了解評鑑過程。本研究欲分析先備概念是否影響評鑑歷程的注意力分配以及使用的認知和後設認知策略，且進一步了解評鑑過程的如何影響評鑑科學解釋的表現。本研究亦嘗試建立眼動行為和認知與後設認知策略間的關聯。本研究採質量混合設計，研究對象為32位大學生，以先備概念測驗得分分為高、低先備知識兩組，進行生物主題的科學解釋評鑑任務。任務過程中，紀錄學習者的眼動行為以分析其在閱讀探究情境與問題、實驗數據及評鑑科學解釋階段的眼動指標。研究者並以眼動提示回溯自陳的方式記錄學習者的思考歷程和判斷結果，分別分析其在評鑑過程中展現的認知與後設認知策略以及科學解釋評鑑表現。除了比較兩組學習者於評鑑表現、眼動行為和認知與後設認知策略之差異，研究者亦進行眼動行為和認知與後設認知策略的相關性分析，輔以以質性分析結果了解評鑑過程如何影響其表現。研究發現，高、低先備知識學習者其評鑑科學解釋之表現無顯著差異。眼動行為分析發現，兩組於各興趣區的第一次注視時間均無顯著差異，但高先備知識學習者在部分興趣區的回視時間、總凝視時間、總凝視次數和掃視次數較低先備知識的學習者高。認知與後設認知策略分析則顯示，兩組在閱讀時皆主要運用認知策略，而在評鑑時則同時運用認知與後設認知策略。比較兩組學習者，高先備知識者在閱讀探究情境時，使用較多有目的的訊息搜尋策略，且在評鑑時使用較多閱讀理解和推理並評鑑策略；而低先備知識者則在閱讀探究情境時使用較多閱讀理解和監測評價策略。眼動指標和認知與後設認知策略的相關分析則發現，兩組的第一次注視時間與閱讀理解策略沒有相關。兩組學習者的評鑑歷程之差異，反映在眼動指標和認知或後設認知策略的相關性結果上：質性分析的結果亦指出，學習者的注意力分配以及使用認知與後設認知策略等歷程之差異，會影響學習者對科學探究活動所能達到的理解狀態，理解狀態較高者，則評鑑表現可能較佳。高先備知識學習者，在關鍵興趣區的總凝視時間和掃視次數與分析應用、搜尋訊息和推理並評鑑等策略呈顯著正相關，理解狀態達到知道自己已理解和知道自己不理解的人數各半；低先備知識學習者在許多興趣區的總凝視時間和掃視次數則與監測評價和監測矛盾策略呈顯著正相關，理解狀態的分析則多為知道自己不理解和未察覺不理解。本研究結果指出，先備知識高、低不僅影響學習者在評鑑過程中的注意力分配，注意力分配和認知與後設認知策略的使用之關聯形式亦不相同。此評鑑歷程的差異會影響學習者對科學探究的理解，進而影響評鑑科學解釋的表現。未來研究評鑑科學解釋也應蒐集過程和結果資料進行三角校正。

關鍵字

先備知識；科學解釋評鑑；眼球追蹤；認知與後設認知策略

並列摘要

Ability to evaluate scientific explanation is one of the main goals in science education. Attention allocation and use of cognitive and metacognitive strategies during the evaluation process may influences the outcomes of scientific explanation evaluation. However, few studies have investigated the evaluation process. This study aims at exploring whether and how prior knowledge influence processes and outcomes of scientific explanation evaluation. Learners’ attention allocation and use of cognitive and metacognitive strategies were analyzed. This study also attempts to establish the association between indices of eye movements and use of cognitive and metacognitive strategies. This study took a mixed-method approach. Thirty-two undergraduates were assigned into either the high- or the low-prior knowledge group according to their score on a biology concept test. The participants’ eye movements during a task of scientific explanation evaluation were recorded. The record of eye-movements was then used as stimuli to elicit the participants’ thinking process retrospectively. Patterns of attention allocation, cognitive and metacognitive strategies, as well as their evaluation performance for the high and the low prior knowledge groups were compared. The correlations between indices of eye-movement and cognitive/metacognitive strategies were analyzed, supplemented with qualitative data, in order to understand how the evaluation process influence the evaluation performance. No significant difference was found between the high and the low prior knowledge group on their performance of scientific explanation. No difference was observed for the first-pass fixation duration. In comparisons with their counter cohorts, however, the high prior knowledge students spent longer time reading and reexamining, as well as more frequently revisiting specific areas of interests. The students in the high prior knowledge group also demonstrated more goal-directed searching while reading the context of inquiry and used more reading comprehension and inferring strategies during evaluation. The low prior knowledge students, on the other hand, used more reading comprehension strategies and more constantly monitored their reading the context of inquiryevaluation process. Findings of the correlation analyses indicated that the frequency of reading comprehension was not correlated with the first-pass fixation duration. For the evaluation process, the differences between the two groups lay in their look-back fixation duration, total fixation duration, and inter-scanning counts on specific areas of interests, as well as in the correlations between these eye-movement indices and frequencies of cognitive/metacognitive strategies at different phases of the evaluation task. For the high prior knowledge group, the aforementioned eye-movement indices were associated with frequencies of analyzing, goal-directed searching, and inferring/evaluating strategies and, therefore, may result in a higher level of understanding about inquiry and better evaluation performance. For the low prior knowledge group, the eye-movement indices were associated with the frequencies of content evaluation and conflict monitoring, which may result in a lower level of understanding about inquiry and the poor evaluation performance. Findings of the study indicated that prior knowledge influences the evaluation process in terms of attention allocation, use of cognitive and metacognitive strategies, as well as the association between eye-movement behavior and strategy use. Future studies aimed to explore the process and outcomes of scientific explanation evaluation should collect these forms of process data for triangulation.