The reliability and validity of a tailored testing procedure based on the simple logistic model was determined for an achievement test in statistics and measurement. The test was administered on a CRT terminal to students from graduate and undergraduate measurement courses. Equivalent form reliability over a one-week interval was found to be 0.595 while KR20 reliability measures for the traditional course exams yielded 0.756 and 0.574. The tailored testing procedure administered 20 items or less while the traditional tests were 55 and 50 items respectively. The tailored test was a valid predictor for the graduate course. (Author)