Criterion-Referenced Interpretation

Criterion-referenced interpretation is the interpretation of a test score as a measure of the knowledge, skills, and abilities an individual or group can demonstrate from a clearly defined content or behavior domain. It is often defined as a contrast to norm-referenced interpretation, where an individual’s score only has meaning when it is compared to other individuals’ scores. Criterion-referenced interpretations are independent of information based on how the average person performs. This entry further describes criterion-referenced interpretation and its uses, then discusses the design and validation of tests that foster criterion-referenced interpretation. The entry concludes with a look at common misconceptions about criterion-referenced interpretation.

Criterion-reference interpreted scores have been used for a variety of decisions, such as monitoring student achievement, evaluating efficacy of instructional programs, granting licensure and certification, planning individual and group instruction, and identifying possible learning disabilities. Tests that are designed to foster criterion-referenced interpretation of scores include Advanced Placement assessments, driver’s license exams, and the Programme for International Student Assessment.

A criterion-referenced interpretation assumes an underlying continuum of content knowledge and [Page 426]behaviors that ranges from none to all encompassing. When the breadth and depth of knowledge and behaviors that comprise the content domain—the criterion—is clearly and completely specified, and a test is constructed with a representative sample of items from the content domain, it is understood that there is a correspondence between an individual’s performance on the test and their ability level on the underlying continuum. Thus, if a test is constructed to foster a criterion-referenced interpretation, the inference can be made that an individual who scores 75% on the test knows and is able to demonstrate individual knowledge of 75% of the content domain.

First outlined by Robert Glaser in his 1963 symposium address to the American Educational Research Association, criterion-referenced interpretation gained popularity in the United States in the 1970s, as the development of theories of measurement and test design refined the distinctions between criterion-referenced and norm-referenced interpretations and their uses. Although it is possible under certain conditions to interpret scores from a single test in reference to both a criterion domain and a norming group, doing so rarely leads to satisfactory interpretations because different score interpretations require different test designs. It is important to note that the nature of test score interpretation (criterion- or norm referenced) is a characteristic of the interpretation as enabled by test design, not the test itself. There is a tendency in the measurement and assessment literature to refer to anything not explicitly norm referenced as criterion referenced. Here, the description of criterion-referenced interpretation is consistent with the original intent.