Criterion-Referenced Assessment

Criterion-referenced assessment stands in contrast to norm-referenced assessment, where grades are distributed according to a predetermined curve and a student's grade depends on their rank relative to the cohort. In a criterion-referenced system, it is theoretically possible for all students to achieve the highest grade if they all meet the criteria, or for no students to achieve it if none meet the standard.

The IB Diploma Programme explicitly identifies itself as a criterion-referenced assessment system. IB assessment criteria define what students must demonstrate at each mark level, and examiners assess student work against these fixed standards. While the IB does set grade boundaries that can shift slightly between sessions (to account for variations in exam difficulty), the fundamental approach is criterion-based: work is judged on its own merits, not relative to other students.

A-Level and IGCSE examinations use a model that combines elements of both criterion-referenced and norm-referenced approaches. While mark schemes define what constitutes a correct answer (criterion-referenced), the process of setting grade boundaries involves statistical analysis of the cohort's performance and expert judgment about maintaining standards over time. The UK regulator Ofqual uses a process called "comparable outcomes" to ensure that grade distributions remain broadly stable across years when the ability of the cohort has not demonstrably changed.

AP exams are designed so that the score scale (1-5) reflects absolute levels of college readiness rather than relative ranking. The College Board states that scores are criterion-referenced, meaning a score of 3 ("Qualified") represents a consistent standard of achievement across years. However, the statistical methods used to equate scores across sessions introduce an element of standardization.

For students, the key practical implication of criterion-referenced assessment is that they should focus on mastering the criteria rather than competing with peers. Studying the assessment criteria, understanding what distinguishes each performance level, and practicing with exemplar responses at different levels are all effective strategies.