概念構圖（concept mapping）是一種經廣泛應用的教學及學習策略，其應用在學科知識測量的可行性，近年來亦頗受學者重視。本研究建立一電腦化概念構圖系統（Concept Mapping Based Learning System）作為學生建立概念圖的環境，並採取Goldsmith, Johson與Acton（1991）的接近性指數（Closensss Index, C指數）以及修改Novak與Gowin（1984）的接近性作為評量學生知識結構的參考指標。 本研究以85名大學生為研究對象，以教育心理學中的「學習理論」相關概念作為材料，探探討概念圖應用於知識結構測量的可行性。結果發現，不同學習成就的學生，在C指數侑N-G分數上都有明顯的差別。此外，當以學期成績為效標變項，以C指數及N-G指數作為預測變項進行逐步迴歸分析，發現C指數對於學生的學期成績有明顯的解釋力，而N-G分數的解釋效果則較弱。由研究結果可知，概念構圖應用於知識結構測量具有某種程度的效度，但記分方法或評量指標的建立，則有改進的空間。

　　　　　Concept mapping has been used in many subject domains as a useful strategy for learning and instruction. Recently, many researchers tried to apply concept mapping to the assessment for structural knowledge. In this study we combined computer technology and concept mapping to build a concept mapping based learning and assessment system. In this system we implemented two indices for knowledge-structure assessment: the Closeness index (C) of Goldsmith, Johnson, & Acton (1991), and the revised Novak & Gowin's (1984) concept mapping scoring rubrics (N-G). eighty-five undergraduates participated in this study. They were asked to represent their content knowledge about the topic of 'learning theories' in educational psychology through the computerized concept mapping system. The results indicated that both the C and N-G indices were significantly different between the students of high- and low achievement groups. In addition, form the results of multiple regression, we found that the C index accounted for a significant amount of the variances of the achievement scores, but the N-G index did not predict significantly over and above the C index. Based on these results, we concluded that concept mapping along with the C and N-G indices can serve as a valid tool for knowledge structure assessment, yet the scoring rubrics could be revised to improve their validity.