Criterion-referenced is typically associated with standardized assessments such as certification and licensure examinations. Simply put, tests which are criterion-referenced are designed to measure candidate performance against a set of predetermined criteria or learning standards. If candidates perform at or above the established expectations (by answering a certain number of questions correctly), they will pass the examination. Passing scores (sometimes called “cut scores”) on certification examinations are typically criterion-referenced. The passing score is set using the educated judgment of a committee of content experts who use psychometric methodology to determine the appropriate number or percentage of correct responses necessary to demonstrate proficiency. Only those who answer enough questions correctly to reach the passing score pass the examination.

This type of approach is in contrast to norm-referenced tests which are typically given in schools. Norm-referenced examinations are tests that are graded on a “curve.” This curve refers to the bell curve which compares the performance of candidates with each other. That is, norm-referenced tests determine whether candidates performed better or worse than a hypothetical average student. This type of approach ensures that a predetermined number of candidates will perform poorly, most will perform in the average range, and a small number of candidates will perform well. With norm-referenced exams, the measurement is made against other test takers and not against a fixed set of criteria.

In other words, with norm-referenced examinations, a predetermined percentage of candidates will pass, but with criterion-referenced examinations, a predetermined number of items must be answered correctly.