Differences in design between norm referenced measures (NRM) and criterion referenced measures (CRM) are reviewed, and some of the procedures proposed on designing and evaluating CRM are examined. Differences in design of NRM and CRM are said to arise from the different purposes that underlie each measure. In addition, there are differences among criterion referenced tests, three cases of which are: (1) where items are sampled from a known universe, (2) where one item constitutes the set in question, and (3) where items are examples of a class of problems or tasks which cannot be well defined. Validation problems in CRM are discussed, and the need for developing new techniques, especially for case (3) CRTs, is pointed out. (DB)