- 學位論文
以CEFR為基礎之華語閱讀測驗系統適性化機制之建立
The foundation of adaptive mechanism for a Chinese reading test system based on CEFR.
摘要
現今華語文能力測驗中，仍以傳統紙筆測驗為主，較少開發線上測驗的系統，國家華語測驗推動工作委員會雖有研發華語文能力測驗(Test Of Proficiency-Huayu，簡稱TOP測驗)線上練習系統，但此系統並不具備適性化之功能。歐洲共同語文參考架構（CEFR）提供了語言能力的具體描述、學習語言之綱要、清楚易懂的能力架構及語言溝通面向等參考細節，在歷經長期的科學研究及廣泛的研議之下，確實提供了清晰的外語學習標準。基於上述動機，本研究目的分述如下： 一、建置一個「以CEFR為基礎並具備適性化機制之華語閱讀測驗系統」，不僅讓受試者可進行施測以了解自己的學習現況，還能提供友善的後端管理介面，方便管理者或教師能快速掌握受試者華語文之程度與能力等級。 二、以實徵資料模擬電腦化適性測驗系統的流程，以評估系統之使用成效。 本研究過程中先以菲律賓僑校811位學生為預試對象，接著從施測結果分析試題參數後，再以此實徵資料模擬電腦化適性測驗系統的流程。根據本研究的結果分析，可以得到以下結論： 一、本研究成功建置電腦化適性測驗系統的使用者介面，讓受試者可以參加測驗並查詢成果，也能讓管理者擁有後端管理的平台。 二、在不同能力估計法的模擬研究中，發現使用此華語文閱讀試題時，整體而言，期望後驗估計法（EAP）優於最大概似估計法（MLE）及最大後驗估計法（MAP），所以建議使用此試題時，使用期望後驗估計法（EAP）。
並列摘要
Traditional paper-and-pencil-based testing is still the primary form of Chinese language proficiency testing today; online test systems are very few. Although the Steering Committee for the Test of Proficiency – Huayu, has developed the TOP (Test of Proficiency – Huayu) Online Mock Test system, it is not adaptive. The Common European Framework of Reference (CEFR) provides specific descriptions of language proficiency, guidelines on language learning, a clear and easy-to-understand proficiency structure, and aspects of language communication. After long-term scientific research and broad discussions, CEFR does provide clear standards for learning a foreign language. Based on the abovementioned motivations, the purposes of this study were as follows: 1. To establish a “CEFR-based, adaptive Chinese reading comprehension test system.” This system would not only allow testees to be cognizant of their learning progress but also provide a user-friendly administrator interface that assists administrators or teachers in quickly checking testees’ levels of Chinese proficiency. 2. To simulate the process of the computerized adaptive test system with empirical data in order to evaluate the effects of the system. First of all, the test system was tested on 811 students from Philippine Chinese School; then, the item parameters were analyzed according to the test results. Lastly, the empirical data were used to simulate the process of the computerized adaptive test system. According to the analysis of the research results, it can be concluded that: 1. A computerized adaptive test system was successfully established in this study. It allows testees to take tests and check test results, as well as provides a back-end management platform for administrators. 2. According to the simulation research on different ability estimation methods, it is found that the expected a posteriori (EAP) estimation is better than the maximum likelihood estimation (MLE) and maximum a posteriori (MAP) estimation, as a whole, in the use of the Chinese reading comprehension test. Therefore, it is suggested that EAP be employed for the test.