When explaining an educational effect using an achievement test, the examination from which difficulty differs to the group from whom an examinee differs will be done. In such a case, although it is thought that the difference in difficulty comes out in the average mark, the method for the difference reflects has not been established yet. In this paper, the model which transformations into another random variable on the random variable is used, and which takes the value of the limited section is based on the idea of drawing the percentage of correct answers in that examination using the difficulty of an examinee's potential academic ability and an examination. The method of estimating the parameter showing the difference in difficulty is raised using this model.