期刊名称：RELIEVE - Revista Electrónica de Investigación y Evaluación Educativa
印刷版ISSN：1134-4032
出版年度：1998
卷号：4
期号：2
DOI：10.7203/relieve.4.2.7210
语种：Spanish
出版社：RELIEVE - E-Journal of Educational Research, Assessment and Evaluation
摘要：Three new item exposure control methods for computerized adaptive tests are introduced. The first one is based on maximum information, and adds a random component to the information provided by each item. The second one directly controls the exposure of each item, and applies only the items with and exposure lower than some arbitrary value. The third method is a mixture of the other two. The three methods, plus maximum information and the Sympson and Hetter method, are compared in a simulation study, with respect to test precision and exposure rate. Finally, we comment some future research lines.
其他摘要：Three new item exposure control methods for computerized adaptive tests are introduced. The first one is based on maximum information, and adds a random component to the information provided by each item. The second one directly controls the exposure of each item, and applies only the items with and exposure lower than some arbitrary value. The third method is a mixture of the other two.The three methods, plus maximum information and the Sympson and Hetter method, are compared in a simulation study, with respect to test precision and exposure rate. Finally, we comment some future research lines.
