摘要
选题是计算机化自适应测验(CAT)测试过程的关键环节,选题策略的目标是要达到较高的测量精度,同时也实现试题曝光率控制及其他测验目标的实现。本文根据选题策略的基本原理和衍生发展,将众多CAT选题策略分为五大选题策略系列:Fisher函数系列、K-LI函数系列、α分层系列、贝叶斯系列、b匹配系列;并根据测验目标(测验精度、试题曝光率控制、内容平衡、多条件约束)对这些选题策略进行了细分,并对CAT选题策略的选择思路进行归纳。
Item selection strategies is the important process of Computerized Adaptive Test（CAT）. Its goal is to achieve high precision, while also taking into account for the item exposure rate control and other conditions of the test target control. Based on the mathematical principles and its main aims of item selection strategies, item selection strategies are divided into five types ： （ 1 ） Fisher information function item selection strategy and its derivative strategies;（2）K -L information function item selection strategy and its derivative strate- gies ; （ 3 ） α - stratified strategy and its derivative strategies; （4） Bayesian strategies and its derivative strategies ; （ 5 ） b - matching strategies and its derivative strategies. These five types series strategies were overviewed and summarized. The author also gives the recom- mendations for item selection strategies under different test cases.
出处
《心理学探新》
CSSCI
2014年第5期446-451,共6页
Psychological Exploration
基金
江西省教育科学“十二五”规划课题(12YB052,0YB254)
江西省社会科学规划项目(13JY47)
国家自然科学基金项目(31260238)
井冈山大学人文社会科学课题(JR10030)
关键词
CAT
选题策略
Fisher函数
K-LI函数
α分层
贝叶斯选题策略
b匹配方法
CAT
item selection strategies
Fisher information function
K -L information function
α -stratified strategy
Bayesian strategies
b - matching strategies
