| In this paper, we investigate repeat test effect of the English college entrance examination reform which is twice a year, analyze the pros and cons of the new policy and construct the IRT model, by using the real test data that is provided by high school teacher. We choose Rasch model in IRT to estimate the parameters. Rasch model can provide the best estimating precision.After we choose the model, we do the model fitting test for the original data, then we estimate the parameter of the student ability and the parameter of the item difficulty by using the condition maximum likelihood estimation and the joint maximum likelihood estimation. At last, we analyze the effect.When analyze the effect, we dissect the test effect to get the time effect, the ability promotion effect, the pattern difference effect, the experience effect, the random effect and the optimal stopping test effect. Then we give the definition of the effect and the mathematics models. At last, we analyze the effect by using the real test data and provide the suggestion of the new English college entrance examination.By studying the condition of no one less, we analyze the ability promotion effect and the pattern difference effect. Then we discover that students’abilities promotion is prominent. The promotion effect is different between the different item pattern. By studying the condition of no one less, we work out the suggestion of the optimal stopping to the second test. By studying the random effect, we calculate the deviation value of the two time test and discover that the mid-ability students are more stable. We contrast the ability difference of three-time test and get the effect of time interval to the college entrance examination policy.The reform of the college entrance examination is just beginning. In this paper, we hope that our research results can provide some directions to the students who will participate the college en-trance examination, decrease the bad effect, get a better score. |