不同BIB設計對測驗等化的影響

Simulate Research on Balanced Incomplete Block

曾玉琳(Yu-Lin Tseng) ； 王暄博(Hsuan-Po Wang) ； 郭伯臣(Bor-Chen Kuo) ； 許天維(Tian-Wen Sheu)

《測驗統計年刊》 13期_下 (2005/12) Pp. 209-229

https://doi.org/10.6773/JRMS.200512.0209

摘要

本研究主要目的在探討使用平衡不完全區塊(balanced incomplete block, BIB)設計在測驗題庫中進行連結等化的效果。本研究探討五種BIB設計，主要控制變項為：1.能力分佈為常態分佈N(0,l)、負偏態分佈、正偏態分佈、雙峰分佈；2.總施測人數約2600人、5460人、約10000人；使用之等化及估計軟體為BILOG-MG。
研究發現使用此五種BIB設計，施測人數為5460人及約10000人且受試者能力分佈為常態分佈或雙峰分佈時，可得到較佳的等化效果。

關鍵字

等化 ； 平衡不完全區塊設計 ； 估計精準度

並列摘要

The purpose of this study is to investigate the effects of five different balanced incomplete block on test equating using the 3 parameter logistic model in BILOG-MG. The impacts of two factors will be discussed: 1. the number of sample, 2. the distributions of ability.
The result shows that the best decision of using balanced incomplete block is sample size exceed 5460 and the distribution of ability is normal or two- kurtosis. When distribution of ability is normal or two-kurtosis distribution, it has good equating effect using balanced incomplete block designs.

並列關鍵字

Equating ； balanced incomplete block design ； accuracy of estimate

參考文獻

王寶墉(1995)。現代測驗理論。臺北市:心理出版社。

Haebara, T.(1980).Equating logistic ability scales by a weighted least squares method.Japanese Psychological Research.22,144-149.

NAEP Mathematics Consensus Project(2001).Mathematics framework for the 1996 and 2000 national assessment of educational progress.National Assessment Governing Board, U.S. Department of Education.

Kolen, M.J.,Brennan, R.J.(1995).Test Equating: Methods and Practices.New York:Springer-Verlag.

van der Linden, W.J.,Veldkamp, B.P.,Carlson, J.E.(2004).Optimizing balanced incomplete block designs for educational assessments.Applied Psychological Measurement.28,317-331.