知識結構的電腦化適性診斷測驗數位平臺應用於國小數學補救教學成效之研究

A Study of Knowledge Structure-Based Computerized Adaptive Test System for Elementary School Students on the Effectiveness of Remedial Math

本研究針對國小數學科補救教學，使用知識結構為基礎的電腦化適性診斷測驗數位平臺，以準實驗設計前、後測蒐集相關資料，比較對照組與實驗組實施數學科補救教學學習成效的差異。以教育部科技化評量篩選測驗低於「絕對通過標準」需進行補救教學的二至六年級學生347人為研究對象，實驗組共101人，對照組共246人。研究結果顯示：一、以「知識結構為基礎的電腦化適性診斷測驗數位平臺，輔之以教師從旁指導」的補救教學模式優於以「傳統課堂講授進行補救教學」模式，且學生進步成績達顯著性差異；二、二年級和五年級實驗組成績表現明顯優於對照組達顯著差異，二至五年級實驗組進步分數幅度皆優於對照組；三、實驗組整體補救教學通過率高於對照組，二年級通過率表現高達七成最佳，年級愈低，補救教學通過率愈高；四、數學學習內容無論是數與量（N）、幾何（S）、代數（A）、統計與機率（D）題型，實驗組的補救教學，學生的學習具成效達顯著差異；五、補救教學學生的自我效能、自主學習能力表現、平臺的使用皆持正向肯定的學習態度。

The purpose of this research is to use the knowledge structure based on adaptive testing system (KSAT), which provides a learning diagnostic report to support remedial mathematics for elementary school students. To evaluate the effectiveness of the KSAT, three hundred and forty-seven students from 2nd to 6th grade who were under the passing criteria of project for implementation of remedial instruction technology-based-test, participated in this study. There were 101 students for the experimental group, and 246 students for the control group. The performances of these two types of remedial instructions were compared by a pre-test/post-test nonequivalent group design. The results demonstrated that: (1) The KSAT is useful for remedial mathematics in elementary school. The remedial learning with teacher assisted guidance on KSAT outperformed another traditional teaching method, and the differentiation of students’ academic achievement were statistically significant. (2) Both the academic achievement and the range of learning improvement in the experimental group from the 2nd and the 5th graders were obviously prior than the control group. (3) The 2nd grade students using the KSAT had the highest passing rate among all groups. As a whole, the passing rate of the experimental group is higher than the control group; the lower the grader is, the higher the passing rate is. (4) The academic achievement of students in remedial mathematic learning has statistically significant improvement no matter on any types of the questions, which includes the number and the quantities (N), geometry (S), algebra (A), statistics and probability (D). (5) All of the learning attitude on students’ self-learning efficiency performance, autodidact performance of the remedial students and the usage of the platform were positively supported.