二级项目题组混合IRT模型及其稳健权重估计

在心理测量和教育测量中,二级项目和题组项目是两类常见的项目类型。由这两种项目混合构成的测试在实践中有着重要的应用。被试在答题时,由于个人的潜在能力和项目难度不匹配,常常会产生异常反应,这些异常反应会影响IRT中潜在特质估计的准确性。仿真实验证明,二级项目题组混合IRT模型的稳健估计方法在出现异常值的情况下,能够比极大似然估计对被试的潜在特质做出更加准确的估计,能够满足实际测试的需求。

Item Weighted Robust Likelihood Method for Ability Estimation Composed of Both Dichotomous and Testlets Items

In the psychological measurement and educational measurement,the dichotomous and testlets items are two common kinds of item types.The test of the mixture of these two kinds of projects has an important application in practice.When the examinees trying to do the items,they often produce aberrant responses because of the potential ability of the individual and the difficulty of the project does not match.These aberrant responses will affect the accuracy of the potential characteristics of IRT.Simulation experiments prove that item weighted robust likelihood method for ability estimation composed of both dichotomous and testlets Items make more accurate estimates,which can fit the needs of the actual test.