Criterion-Referenced Assessment

What is Criterion-Referenced Assessment?

Although criterion-referenced testing is the term most commonly used, educators do not agree that it is the best term to describe the assessment method in which score interpretation is based on a predetermined, set criterian of performance. During the past 30 years, terms such as competency tests, mastery tests, domain-referenced measurement, objectivesreferenced assessment, curriculum-based assessment, and performance or direct assessment have been used in place of criterion-referenced assessment. More recently, as states and school systems have adopted learning and curriculum standards that describe what students should know and be able to do in core subjects at each grade level, they have developed a variation of criterion-referenced assessment referred to as “standards-referenced testing”or“standardsbased assessment.” Furthermore, the broader term assessment seems to be replacing the term testing.

Criterion-and Norm-Referenced Testing (Assessment)

As originally defined by Glaser (1963) in American Psychologist, criterion-referenced tests “depend upon an absolute standard of quality,” while norm-referenced tests “depend upon a relative standard.” Since Glaser's original article, it has become common to label tests as criterion-referenced or norm-referenced. However, these labels actually apply to the interpretation of the examinee's performance as expressed through test scores rather than the testing instruments themselves.

This shift in emphasis from test type to test score interpretation is captured by Popham in his 2002 edition of Classroom Assessment: What Teachers Need to Know, in which he says “…a criterion-referenced interpretation is an absolute interpretation because it hinges on the extent to which the criterion assessment domain represented by the test is actually possessed by the student. Once an assessment domain is defined, the student's test performance can be interpreted according to the degree to which the domain has been mastered” (p. 111).

Linn and Gronlund (2000) propose that criterionreferenced and norm-referenced assessments should be “viewed as ends of a continuum, rather than as a clear-cut dichotomy” (p. 44). Typically, criterion-referenced tests focus on a delimited and clearly specified achievement domain (i.e., objectives and/or learning standards) and emphasize what learning tasks students can and cannot perform. In developing criterion-referenced tests, a large pool of test items are constructed or selected that match the learning task. Their level of difficulty is determined by the difficulty of the task. Test items are eliminated only from the pool of items if they are judged not to measure the learning task or objective.

An important consideration in test item selection for published norm-referenced tests is the statistical properties of the individual items. Items of moderate difficulty and high discriminating power are most likely to be included because they will yield distributions of test scores that enable educators to distinguish performances among individuals. To interpret a person's performance, the performance is compared to the performance of other examinees with similar characteristics.

...