Criterion-referenced test requires a well defined domain and a cutoff score. To'set a cutoff score is a controversial issue for CRT. Establishing cutoff score has traditionaly been preformed in two phases:obtaining cutoff scores based on expert's subjective judgment and establishing a cutoff score by a statstical method. Mills and Melican(1988)c1assfied the all proposed methods for setting cutoff scores with four categories. Four categories are normative method, aboslute method based on evaluation of examinees, absolute method based on evaluation of the test, and compromise method. This article introduces several methods for setting a cutoff score. In a category of the absolute method based on evaluation of the test, there are the Nedelsky method, the Angoff method, the Jaeger method, and the Ebel method. The Nedelsky method has a drawback which the cutoff score based on the Nedelsky method is generally lower than these based on the other methods. The cutoff score based on the Nedelsky method is determined by summing the reciprocal of the number of options not eliminated by a judge across items. Therefore, there is no probability of correcting an answer between 1/2 and 1 for each item. This study modified the Nedelsky method to set a cutoff score. A cutoff score is determined by summing the reciprocal of the number of options which are considered as incorrect options by judge cross items. This modified method for the Nedelsky method is called as "the absolute method based on ability of distracting incorrect options". To investigate a validity of the method based on ability of distracting incorrect options, thirty high school mathematic teachers in Seoul were selected randomly from a sampling frame of the "List of Korea School". These teacher were asked to analyze each item of the "Mathematic I test of the 1990 College Entrance Examination" based on the Angoff method, the Jaeger method, the Ebel method, and the absolute method based on ability of distracting incorredct options. The eleven teachers' responses were analyzed for the study. This study found that the cufoff score of the Nedelsky method was differ from these of the other methdos significantly. The cutoff score of the absolute method based on ability of distracting incorrect options was not differ from these of the other methods except the Nedelsky method. Therefore, the method based on ability of distracting incorrect options is recommendable for setting cutoff scores. Additional finding of the study was that the "Mathematic I test of the 1990 College Entrance Examination" was not easy. Mastery learner for high school mathematic curriculm would get 61.65 for the test.