RÉSUMÉ
In order to compare a group of examinees to other group, the tests taken by two groups of examinees should be equivalent. The first step of the equating the test is to make anchor items. In medical schools, the students prepare the examination through the thorough review of the test items of previous year. So it is said that the reuse of the same items could be undesirable. The purpose of this study is to find out response variations to the same items between two consecutive grades when the items are reused. The senior classes of a medical school are sampled and the test items of graduation examination was analysed. On the basis of item difficulty(item's p-value) and discrimination index, we selected 35 items. Next year, we reused those items to the same examination to the senior class of that year, and the result was analysed. Among those 35 items, 14 items were modified slightly. The averages of the item difficulty and discrimination index on the previous examination were 0.49 and 0.20 which were in the desirable ranges. But at the next year these data were worsened to 0.84 and 0.10 respectively. This trends were not different in the slightly modified items. And there was no significant differences among the item groups classified on the basis of the level of knowledge. We could ascertain that if a previously used item used again to a test, the item difficulty will increase(become easy) and discrimination index will decrease even though minor modification is done.