The purpose of this study is to examine an option-weighting method as it affects pass-fail decisions in formative and summative evaluation of student achievement for instructional units, certification, advancement, licensure, admissions, placement, and selection. A database was constructed using high school achievement test data where a unidimensional domain was established and short tests simulated to test the validity of the option weighting procedure against the standard procedure for scoring test items. The weight for each option was its discrimination index. It has long since been established that option weighting positively affects estimates of reliability but has unclear effects on predictive validity. In this study, reliability was directly computed, and option weighting again proved effective in increasing reliability. More importantly, option weighting applied to a pass-fail decision-making situation provided a substantial improvement over regular scoring for tests which were constructed by simple random sampling from a well-defined domain. The advantages for option weighting were greatest when the test contained 10 items; as test length increased, the advantage of option weighting seemed to lessen. (BW)