Three kinds of classificatory decisions that might be made using criterion-referenced tests (CRTs) are described, and methods to appraise the validity of each are subsequently discussed. Decisions that entail predictive, descriptive, and evaluative classifications comprised the three kinds of decisions described. Predictive classifications entail assigning individuals who take a test to categories that are intended to predict their status on a criterion variable. It is suggested that these decisions can be appraised for validity using forms of the familiar criterion-related validity study. Considered as examples of descriptive decisions are the mastery and competency decisions that are commonly made using CRTs. Evaluative classification decisions would be made when judges set a passing score regarded as desirable to require. When these performance classifications reflect judges' intrinsic valuations of different levels of test performance, it is argued that the validity of these classfications cannot be shown. However, when these performance classifications reflect judges' instrumental valuations of different levels of test performance, their validity should be shown by criterion-related validity studies. (Author/GK)