Technical considerations associated with item selection and reliability assessment are considered in relation to criterion-referenced tests constructed to provide group information. The purpose is to emphasize test building and the evaluation of test scores in program evaluation studies. It is stressed that an evaluator employ a performance or achievement instrument sensitive enough to reflect group/individual ability in terms of the specific goals of the program. Criterion-referenced tests are designed for this purpose. Two steps in criterion referenced test development are considered: the approach to item selection, and the assessment of score reliability. Methods for handling them in preparing and using criterion-referenced tests in program evaluation studies are offered. Norm-referenced and criterion-referenced testing within a program evaluation context are compared, and a model for developing/validating criterion-referenced tests is introduced. (Author/GK)