Two aspects of criterion referenced testing are discussed: cutting scores and test length. Several practices in determining passing scores are enumerated: (1) setting passing scores so that a predetermined percent of students pass; (2) inspecting each test item to determine how important it is that it be answered correctly; (3) determining the educational consequences of setting the cutting score too high or too low; (4) examining the psychological and financial costs of a high or low passing score; and (5) adjusting the score to allow for measurement error. Test length is based on determining the size of a sample of problems needed to acquire an estimate having a specified level of accuracy. Approaches to determining test length are classical test theory, the binomial model, and sequential models. Two functions of testing programs are to assess the competency of a student in regard to specified instructional objectives and to evaluate program effectiveness. Procedures and tables are given to help determine how many test items are needed to assess individual achievement and how many students should be tested for program evaluation. (MH)