沿用曝光控制因子的同时,基于抽样原理,引入区分度分布因子,按区分度的分布情况来选取测验中的项目.以lna-N(0,1)为例,Monte Carlo模拟结果表明:该方法在估计精度、效率和安全性等指标上表现得比较优异.
The exposure-control factors has been followed, at the same time, a discrimination-distribution factor has been introduced. According to the distribution of discrimination, the new item selectisn srategy to balance item usage exposure rate has been proposed. Suppose that lna -N(0, 1 )and according to Monte Carlo simulation, the results show that the new approach has more preferably performance comparing with other approaches on several assessment indexes.