以眼球追蹤技術分析國中學生生物試題之閱讀與理解

Abstract

標準化紙筆測驗是台灣學生的學習及升學過程中，主要的評量工具，因此，了解學生在解決標準化試題時的閱讀歷程，以及此歷程與解題成就間是否存在著關係，可以讓我們進一步推論學生的問題解決策略。本研究根據布魯姆認知分類修訂版架構，分析生物基測試題，選出代表題型後，進行成就測驗，在學生解題同時，利用眼球追蹤技術，紀錄國一學生解決試題時的閱讀歷程。眼球追蹤資料經過統整後，使用SPSS 19軟體進行歷程分析。此外，依據試題測驗結果的分為高低兩組，進行描述性統計、成對樣本T考驗、同質性檢定與單因子變異數分析（ANOVA）。 結果發現，學生在解決布魯姆認知分類架構下較高階或有圖的試題時，需花費較多的凝視時間及總凝視點數，這表示學生需花費較長的訊息處理歷程來解決這些問題，進一步的相關分析發現，除了了解-敘述題型外，學生在不同題型上的測驗成績與閱讀歷程大致呈正相關，顯示訊息處理時間長短會影響作答結果。而高低組學生的差一分析指出，高分組學生在解題時會將較多的注意力分配於試題中解題資訊所在；此外，高分組的學生通常也會有較多的交互閱讀及回視次數，這表示高分組學生在解決試題時，有較多整合題目所給予的資訊的過程，而這些過程皆可以幫助學生正確解決問題。Standardized paper and pencil tests are the main tools for assessing students’ learning in Taiwan. For this reason, it’s worth of a further study for the relationship between reading process when students are solving the standardized test problems and their problem-solving achievements. In this study, we used the eye tracking method to collect the 7th grade students’ eye movement patterns when they were solving the national standardized examination on the topic of biology. Stduents’ text performances were also calculated and divided in to high and low achievement groups. We used SPSS to conduct statistical analyses on the eye movement data, including descriptive statistics, Paired-Samples T Test. Moreover, the test of homogeneity, and one-way ANOVA were performed to find the associations, if any, between eye movement data and the outcomes of standarized examination. The study found that in general students displayed higher fixation time and Number of fixations on test items which were either classified as the higher-order items according to the Revised Bloom's Taxonomy, or items illustrated with graphics. The result indicated that students needed more information processing time for these types of test items. In addition to the hierarchical orders, problem-solving related information also affected the distributions of students' attention. Comparisons between students of high and low performances showed that high-performance students focused more on where the problem-solving related information was. Furthermore, the high-performance students displayed higher frequencies of Inter-scanning and regression. Such findings suggested that high-performance students were better at integrating test information located at different parts of the test item.