Primary Construction Of Mathematics Achievement Computerized Adaptive Test For Grade 4th

The evaluation on the performance of the primary students is the essential part of the primary education and the key in the reformation of the examination in the primary schools as well. The current evaluation of the students' performance mainly involves two aspects: the goal of the fundamental development and the aim of study on the subjects. The evaluation on the performance of the study of certain subject is primarily based on the curriculum criterion of the various courses. The achievement test designed according to the curriculum criterion is to measure to what degree students master the concepts, the principle and the knowledge in certain fields. Additionally, all educationists are looking for a kind of method which will evaluate the performance of students scientifically and fairly. In order to improve the evaluation on the performance of students and help teachers measure the degree of students' mastery on the mathematics so as to give each student a separate test, the author, borrowing the experience from America and Europe on exam of fundamental education and under the instruction of the item response theory and the computer technology , designed the criterion referenced mathematics achievement Computerized Adaptive test (CAT) to provided a subsidiary test for the evaluation of the mathematics performance of Grade 4. To ensure the efficiency and accuracy, two-way specification table and testspecification, before the experiment, were designed, in terms of which the corresponding items were carefully determined. Adopting the form of the single choice and the basic score 0 or 1, the items finally developed four test papers, each of which involves 50 items. The Bejar method was carried out on the collected data and all the results were in accordance with the request of unidimensionality .When evaluated by ANOTE software, or so 90% data fitted Three-Parameter Logistic Model. After the anises of checkout result of model and preferences, only 141 items were taken in the item storage . The cut-off score was determined by the experience estimation method, and it is corresponding capacity value was given by ANOTE software . In CAT the value was looked as the pass standard of student capacity. During the process of compiling the CAT ,Maximal likelihood estimators and the Newton Raphson method (N-R) were used to estimate the student capacity , and treated test information =25 as the test end standard , which was accorded with the requirements of the test precision.