Estimating Raw Items Parameter In CAT

Recently, with the development of computer technology and increasing needs of individual learning, CAT causes more and more concern .However, the security of adaptive test is also confronted with some new challenges. Lot's of factors, for example items used repeatedly and shared among the candidates ,affect the safety, effectiveness and fairness of the adaptive test .Therefore ,it is a very important issue in the research on computer adaptive test that build a large-scale, high-quality CAT item bank.At present, the items of CAT are generally constructed by experts, and then test takers are employed to acquire their responses to each item, estimation of item parameters and verification of FIT will be done. High-quality item and corresponding item parameters merge into the bank together.However, several problems go with this method. First, a great number of test takers are needed, which demands time and energy with high cost. Second, the safety of the items cannot be guaranteed in the pilot test. Third, techniques of equating need to be employed to construct a good item bank. Complicated techniques cannot be ensured the safety. Especially the anchor items repeatedly used in the test equating will get known more easily.If raw items can be inserted in the process of CAT and the item parameters are estimated at the same time, it will be significant for the construction of CAT item bank. Now, it has not seen such kind of report in this field either in the national magazines or the foreign researches. Thus, the present research is to study the insertion of raw items and their parameters'estimation. Simulation of Monte Carlo has been employed to estimate the parameters of raw items with the one-parameter Logistic and two-parameter Logistic models in the study and good result has been gained.