透過您的圖書館登入
IP:18.118.31.247
  • 學位論文

具試題曝光率控管之IRT適性測驗演算法

The algorithm of IRT CAT system with item exposure rate controlling The algorithm of IRT CAT system with item exposure rate controlling

指導教授 : 郭伯臣
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


摘要 電腦適性測驗(CAT)之題庫建立不易,如不能有效控制題庫題目曝光率,當題目過度曝光,受測者則容易施測到相同的題目而影響到測驗的安全性與公平性。 以能力分佈為基礎之SHC曝光率控管法-ASHC(陳昇座,2007),依據學生能力分布給予不同的曝光控制,人數較多的能力區間給予較嚴格的曝光率控制,人數較少的能力區間給予較寬鬆的曝光率控制,藉此改善SHC(Chang, Ansley, & Twu, 2002)曝光率控管法中受試者能力估計精準度較差的問題,並且維持SHC 曝光率控管法對試題曝光率的良好控管特性,然而該方法必須事先獲得學生能力布情形以推算各能力區間曝光率控制值,增加實際應用的困難,且無法避免施測前須迭代模擬曝光參數的過程,一旦施測情境與原先迭代曝光參數時的情境不同時,曝光參數便無法有效控管試題曝光率,此外在偏態的情境下,ASHC的最大曝光率有偏高的現象。 本研究提出將線上控管法加入ASHC中-ASHC online with freeze(ASHCOF),以避免迭代模擬曝光參數的困擾,讓施測成效不受施測情境異動所影響,研究發現,ASHCOF能有效改善ASHC曝光率控管,在能力值估測維持良好的效能,並且在曝光率控管較寬鬆及題庫較大的情況下優於其他線上控管法。

並列摘要


Abstract In Computerized Adaptive Testing (CAT), it costs a lot to build items. When the items are over exposure, the examinees are easy to take the same items. The security and equity of the test are questions. The ability distribution based SHC(ASHC)Procedure was proposed by Sheng- Tso Chen. They fund that ASHC improve measurement precision and maintain controlling of items exposure well by comparing with SHC Procedure. However ASHC procedure have to get the ability distribution of the examinees before testing. Item exposure parameters of ASHC can be obtained only through time-consuming simulations prior to operational. Besides, ASHC’s maximum item exposure rate is too large in some conditions. The purposes of this study are to implement ASHC online with freeze procedure, and to compare ASHCOF with others online approach. The results showed that ASHCOF can improve controlling of items exposure and maintain measurement precision well by comparing with ASHC Procedure. In case of lower exposure controlling and larger item pool, ASHCOF is a better choice of the online approach.

並列關鍵字

無資料

參考文獻


朱怡君(2005)。a-分層電腦適性測驗之曝光率控管,國立中正大學心理研究所碩士論文,未出版,嘉義縣。
吳玫玲(2006)。電腦適性測驗線上曝光率控管之研究。國立中正大學心理研究所碩士論文,未出版,嘉義縣。
Baker, F. B., & Kim, S. H.(2004). Item Response Theory : Parameter EstimationTechniques. Basel, N. Y. : Marcel Dekker, Inc.
Chang, H. & Ying, Z. (1999). a-stratified multistage computerized adaptive testing.
Applied Psychological Measurement, 23, 211-222.

延伸閱讀