Item Scoring Model Based On Item Node And Its Application

By discussing the defects of Classical Test Theory (CTT), this thesis firstly introduces the basic contents of Item Response Theory (IRT) which include its hypotheses, models and superiorities.Then, the definition of item node is given by the steps of answering a math item. On the assumption that the probability of a correct response on an item node for a subject can be fit by the two-parameter Logistic Model (2PL), three polytomous scoring models, one of which is as the same as Grade Response Model (GRM), are proposed for three types of items. Meanwhile, the three-parameter Logistic model (3PL) is derived by2PL.After the models are proposed, the thesis further discusses procedures to estimate parameters. In this part, an indicator function for a item node is defined and the likelihood equation for each item node is then given. Then, the steps to estimate item node and ability parameters simultaneously are illustrated.In the last part, an application of these models and estimation procedure is given by an example.