中文摘要：

项目反应理论中4PL项目反应模型是在3PL项目反应模型基础上发展起来的,它进一步对模型进行优化,加入失误参数,即引进一条上渐近线,使得估计结果更加准确。传统的项目反应模型往往忽略题目之间的相依性,所以为了使估计更加符合实际,将4PL项目反应模型与题组反应模型结合起来,提出了4PL题组反应模型。将这个模型运算出的结果与4PL项目反应模型的结果相比较,对比结果表明：题组效应不可忽略,2组参数估计结果差异显著。其中,针对4PL项目反应模型和4PL题组反应模型的参数数据,采用Win BUGS软件进行处理。

英文摘要：

The four parameters Logistic（ 4PL） item response model in Item Response Theory is on the basis of 3PL item response model to further optimize the model,joining an error parameter. Its meaning is to join an upper asymptote,and it makes the estimated results more accurately. The traditional item response models ignore the dependence between the items. In order to make the estimated results practical,we combine 4PL item response model with testlet response model. The 4PL testlet response model is proposed in this case. Then we can compare the 4PL testlet response model with 4PL item response model. The results show that we can not ignore testlet random effect and the parameter estimation of the two models are very different. It is a good idea to process data by Win BUGS software.