Characteristics that distinguish criterion-referenced tests from their norm-referenced counterparts are discussed, including: the purposes that they are designed to serve; the characteristics of the types of items that they contain; and the manner in which they are developed. More specifically, the distinguishing characteristics include: reference for measurement; information obtained; homogeneity of item difficulty; standards of performance; distribution of outcomes; and item format. A critical difference is the reference for measurement. For criterion-referenced tests, the reference of measurement is the existence of the behavior of interest. The criterion-referenced test assesses whether or not the individual possesses the skill. The qualitative distinction about performance is optimal for assessing specific or minimum competencies. The reference for measurement for norm-referenced tests is the distribution of outcomes in a representative normative sample. The question asked is how the individual performs relative to others from the reference population. Performance standards are arbitrarily imposed on the norm-referenced measure. Normative tests are optimal when the entire range of knowledge or skill is of interest. These differences are summarized in a table. (SLD)