頁籤選單縮合
題 名 | 電腦化適性測驗題庫之品質管理策略=The Maintaining Strategies on the Quality of Item Bank in Computerized Adaptive Testing |
---|---|
作 者 | 何榮桂; 陳麗如; | 書刊名 | 國家科學委員會研究彙刊. 人文及社會科學 |
卷 期 | 9:4 1999.10[民88.10] |
頁 次 | 頁630-652 |
分類號 | 521.3 |
關鍵詞 | 品質管理; 項目反應理論; 電腦化適性測驗; 題庫; Item response theory; IRT; Computerized adaptive testing; CAT; Item bank; Quality management; |
語 文 | 中文(Chinese) |
中文摘要 | 本研究是一項新的嘗試;旨在運用品質管理的技術,結合項目反應理論以及電腦 化適性測驗,建立 IRT-Based 題庫品質管理模式, 提供有效的品質管理策略,以篩選出品 質不佳試題,減輕題庫維護工作之負荷,維持精確測量之品質。研究分為三步驟,首先,依 品質管理模式規劃 IRT-Based 題庫管理流程, 並採用品質管理技術, 以 Urry、Baker 及 MicroCAT 等所建議之試題參數範圍當作管制條件(分別以 U 型、 B 型及 M 型表示之), 配合「寬鬆」、「中度」及「嚴格」等三種不同管制程度,形成九種題庫品質管理策略;其 次,實際開發品質評估系統以及模擬實驗;最後,再以實驗結果評估各題庫品質管理策略的 效能。 研究結果發現,以「 B 型 - 寬鬆」策略較能有效篩選 IRT-Based 題庫中試題訊息 量較低之不良試題。倘若採用嚴格或中度品質管理策略,皆易發生誤將使用頻率頗高之試題 評估為不合格的現象。此外,在真實題庫之應用上,建議分別修正 Baker ( B 型)管制條 件中 a 參數範圍之上限至 3.0, 以及 MicroCAT ( M 型)管制條件中 C 參數之下限至零 ,才能有較佳的品質管理效果。 在模擬題庫方面,則需調整試題缺點數計算公式中 a 參數 之權數,才能提高品質管理策略之有效性。本研究結果可做為電腦化適性測驗及網路化適性 測驗之動態題庫設計、 開發及維護之參考。 |
英文摘要 | This initial work aims to apply the techniques of quality management, item response theory and computerized adaptive testing theory to develop the strategy of maintaining the IRT-based item banks. The implementation of the work consists of three stages. First, the logical procedure of maintaining IRT-based item banks base on the technology of quality management, adopting three conditions (Urry's, Baker's and MicroCAT's suggestions) and three degrees (loose, medium and strict) of quality control to elaborate nine strategies were constructed. Second, the quality evaluating system was designed and implemented, and some simulation experimentations were conducted. Finally, the effect of the procedure and nine quality control strategies were evaluated. The results indicated that the "B type-loose" strategy is better than the others in screening the items with lower item information from IRT-based item banks. When adopting "U type-strict", "B type-strict" and "M type-strict" strategies to maintain item banks, some useful items misjudged. Because the criterion is too strict. Similarly, the similar result appears when the medium strategies were applied. Moreover, to modify the upper bound of the a parameter from 2.0 to 3.0 in Baker's suggestions (B type) and the lower bound of the c parameter form 0.05 to 0 in MicroCAT's suggestions (M type), it will make better efficiency in maintaining work of real item banks. Adjusting the weight of the a parameter in the item-defect formula will increase the utility of quality control strategy in makeup item banks. The findings of this work will provide the designing, developing, implementing and maintaining guideline of dynamic item bank for computerized adaptive testing and even networking adaptive testing. |
本系統中英文摘要資訊取自各篇刊載內容。