A new standard setting approach is introduced, called the cognitive components approach. Like the Angoff method, the cognitive components method generates minimum pass levels (MPLs) for each item. In both approaches, the item MPLs are summed for each judge, then averaged across judges to yield the standard. In the cognitive components approach, items must be decomposed into nonoverlapping cognitive components that may be thought of as specific skills or knowledge required for a correct response to an item. The method was studied with 12 judges, all third- or fourth-grade teachers who judged a sample of 2,500 students from a third-grade state criterion-referenced mathematics test. Teachers also used the Angoff method to set standards for these results. The most surprising finding of the study was the similarity between the two sets of results. Results from the cognitive component method resembled those from the Angoff method in the range and standard deviation of the recommended standards, as well as in the final standard itself. Interjudge variability was considerably smaller for the cognitive components responses than for the Angoff responses. Some of the validity concerns that may be raised by the cognitive components method are discussed. Additional studies are necessary to support the use of the method in setting standards, and the method is probably only useful when test items lend themselves to decomposition into subtasks. (Contains 3 tables and 10 references.) (SLD)