Norm-Referenced and Criterion-Referenced Tests : Theory and Practice

Abstract

Norm-referenced and criterion-referenced tests are both achievement or proficiency tests, each of which has a different theoretical background. Construction of these tests with incorrect applications of theoretical backgrounds will lead to incorrect interpretation of raw score, resulting in huge negative effects on students and all other stakeholders. Both approaches have long been developing for more than half a century, therefore, in this article they will be named “theories”.

            Norm-referenced test theory explains that in general, the test domain is large and heterogeneous.                                         A test measuring such a domain is only a sample of tasks. A raw score is not a true score and not dependable as a true score.  It cannot be interpreted as a percent correct as a true score. It is, then, interpreted through norm referencing. While the criterion-referenced test theory explains that in some cases, the test domain is small and homogeneous.  A small sample of tasks can represent the whole domain, just like a spoon of soup can represent the whole soup in a pot. Therefore, a raw score is a virtual true score and dependable as a true score. It, then, can be interpreted in terms of a percent correct referencing directly to the test domain.