A criterion-referenced test measures a student’s performance against a fixed, predetermined standard or set of learning objectives, rather than comparing that student’s score to the performance of other test-takers. A driving test is a commonly cited example, since passing depends on meeting a fixed standard of driving skill set by a licensing authority rather than outperforming other applicants tested that day.
Design and scoring
Criterion-referenced tests are typically built directly from a specific set of learning standards or objectives, with each test item designed to measure whether a student has mastered a particular skill or piece of knowledge named in those standards, and results are usually reported as pass or fail, or as a score against a cut point, rather than as a percentile rank. Many state-level standardized tests used in K-12 education are criterion-referenced, designed to report whether a student met, exceeded, or fell below a defined proficiency standard tied to the state’s academic curriculum.
Comparison with norm-referenced testing
Criterion-referenced tests differ from norm-referenced tests, which report a student’s performance relative to a comparison group rather than against a fixed standard, and the two approaches answer different practical questions: a criterion-referenced result tells a parent whether a child has met a specific grade-level expectation, while a norm-referenced result tells a parent how a child compares with peers nationally. Some testing programs report both types of scores from the same assessment, giving educators both a measure of standards mastery and a sense of how a student’s performance compares with a broader population.
Setting the cut score
A key step in building a criterion-referenced test is deciding exactly where the passing cut score falls, a process that typically involves panels of content experts and educators reviewing test items and judging how a borderline student who just barely meets the standard would be expected to perform, a method often called standard setting. This process is inherently more judgment-based than it might first appear, since the underlying standard itself, such as grade-level reading proficiency, is a human-defined target rather than a naturally occurring boundary in the data.
Professional licensing exams, such as those required to become a nurse or an electrician, are typically criterion-referenced for the same reason driving tests are, since the public interest lies in confirming a candidate has met a fixed minimum competency standard rather than in ranking candidates against each other.
Advantages for instructional feedback
Because criterion-referenced results are tied directly to specific, named skills rather than a relative ranking, teachers can use them to identify precisely which standards a class as a whole has or has not mastered, informing decisions about which topics need re-teaching before moving forward. This diagnostic clarity is one reason many states built their primary accountability testing systems around a criterion-referenced model even while also reporting some norm-referenced information for context.
Use in curriculum-aligned classroom testing
Beyond large state assessments, many classroom-level unit tests are also effectively criterion-referenced, since a teacher typically designs a test to measure whether students learned the specific content covered in that unit rather than to rank students against a broader population, making criterion-referenced logic the more common approach at the classroom level even when teachers do not use the formal term.
Reporting formats for parents
States and districts using criterion-referenced assessments typically report results to parents using achievement level labels, such as below basic, basic, proficient, and advanced, each tied to a description of what a student at that level can typically do, which is intended to make the fixed-standard nature of the score more intuitive to a parent than a raw numeric score alone would be.
Historical development of the term
The term criterion-referenced measurement was introduced into educational psychology literature in the 1960s, around the same period mastery learning and other standards-focused reforms were gaining attention, reflecting a broader shift in educational measurement away from an earlier emphasis on ranking students along a normal distribution and toward measuring whether specific, explicitly defined learning objectives had been met.
Workplace certification programs, such as those for project management or information technology credentials, generally rely on criterion-referenced exams as well, since employers and certifying bodies care primarily about whether a candidate meets a fixed competency bar rather than how that candidate compares with everyone else who happened to test on the same day.
Standard setting panels in practice
A standard-setting panel convened to establish a criterion-referenced cut score typically includes classroom teachers, curriculum specialists, and sometimes outside psychometric consultants, who review actual test items and discuss in detail how a student right at the boundary of proficiency would be expected to perform on each one, a methodical but ultimately judgment-dependent process that test publishers document carefully to defend the resulting cut score against later legal or policy challenges.
Large-scale international assessments that compare academic performance across countries generally rely on different measurement approaches than a typical national criterion-referenced test, since comparing fixed national standards directly across countries with different curricula raises its own methodological challenges.
Vocational and trade certification bodies frequently publish detailed criterion-referenced blueprints describing exactly which skills a candidate must demonstrate, giving training programs clear guidance on what to teach well before candidates ever sit the actual exam.
Because a criterion-referenced test’s cut score is fixed rather than relative, it is theoretically possible for every single test-taker in a given year to pass, or for every test-taker to fail, an outcome that would be statistically almost impossible on a norm-referenced test by design.
FAQs
Editorially reviewed by Specialty Digest Editorial TeamLast reviewed November 2, 2026Researched and drafted with AI assistanceReport an issue