Formal assessment is well established in the military and government for applications such as initial selection, promotion, and end-of-course training. For end-of-course assessments of lengthy and/or critical training, it is particularly important for leadership to be confident in the passing score required. Qualified people must not be excluded from passing, and unqualified people should not pass. Yet currently, required passing scores for criterionreferenced tests are often set using arbitrary methods. Although such methods may take into account the criticality of the content overall, they do not use a detailed enough description of job performance requirements to establish “minimally acceptable levels.â€� This paper describes the importance of establishing a rational passing score, or cut score, and several ways of establishing cut scores, focusing on the Modified Angoff (MA) method. This widely used conjectural method has been adjudicated in the courts and is therefore considered defensible. The MA method features a group of informed judges independently estimating what proportion of minimally qualified test takers will correctly answer each test question. Advantages and disadvantages of the method are discussed, as well as factors in its successful application. The authors have employed the MA method for several years on behalf of the Veterans Benefits Administration (VBA) Skills Certification program, a system of Congressionally mandated, high-stakes certification tests. Although the MA method may be applied to a variety of assessment tests and formats, the VBA tests are comprised of multiple choice and similar test item formats, and are delivered online. The process used to collect judges' estimates is discussed, including the frame-of-reference training provided, the technology supporting the intake of ratings, and the computation of cut scores for these tests. Compared with arbitrary methods, the MA method provides greater assurance that those who pass are, indeed, qualified to pass.