Formative Assessment Examples That Actually Measure Understanding

Most lists of formative assessment examples are really lists of recall checks with a friendlier name. A poll, a quiz, an exit slip: each one asks a student to recognize or retrieve an answer, then calls the result a measure of understanding. It rarely is.

Formative assessment was never meant to be a smaller, more frequent version of a test. Foundational research frames it as a process where evidence of student understanding is continuously collected, interpreted, and used to guide teaching, not a mechanism for measuring achievement after the fact. That distinction, process versus checkpoint, is where most classroom tools quietly fail.

What Counts as a Formative Assessment Example

The category is broad by design. Cold calling, think pair share, one minute papers, concept maps, low stakes quizzes, exit tickets, and peer review all qualify because they happen during learning rather than after it. A mixed method study of formative assessment practices in higher education, published in Frontiers in Education, found that self, peer, and group assessment strategies were among the most common formats used with pre-service professionals across university programs.

Breadth is not the problem. The problem is that most of these examples, as typically implemented, measure whether a student can produce a correct answer under light pressure. They do not measure whether the student understands why the answer is correct, or whether that understanding would survive an unfamiliar question.

Where Common Examples Fall Short

Take the exit ticket. A student answers one or two questions on the way out of class, and the instructor scans results and adjusts tomorrow's lesson. This is formative in structure, since it happens mid process and generates feedback. But an exit ticket built on a short factual answer measures the same thing a multiple choice question measures: retrieval under a different label. Why Multiple Choice Tests Measure Memory, Not Understanding

Polls and quick checks have the same ceiling. They can flag that a student got a question wrong. They almost never explain why, because a wrong answer on a four option question compresses a rich, structured misunderstanding into a single incorrect click. The instructor sees the symptom, not the misconception underneath it.

This is not a small gap. A 2025 learning analytics study in the Journal of Computer Assisted Learning found that how a formative check is operationalized, not just how often it occurs, shapes both student engagement and its power to predict academic achievement. A well-timed check with a shallow instrument is still a shallow instrument.

Examples That Actually Surface Understanding

The formats that get closer to real measurement share one trait: they require the student to generate an explanation, not select one.

Concept mapping asks students to draw the relationships between ideas, which exposes whether they understand structure, not just terminology. Peer teaching, where one student explains a concept to another, has decades of support behind it. Research on the protégé effect shows that learning by teaching helps students assess their own understanding and identify knowledge gaps, and that active teaching produces greater gains than simply preparing to teach. Think-alouds, where a student narrates their reasoning through a problem in real time, work on the same mechanism.

The underlying principle is called the self-explanation effect. A meta-analysis of 64 studies covering roughly 6,000 participants, summarized by the British Psychological Society, found that generating an explanation is consistently more effective for learning than simply receiving one, across most subject areas and for both conceptual and procedural knowledge. Explaining does not just demonstrate understanding after the fact. It builds and repairs it in the process.

This is also why explaining is a more sensitive diagnostic than answering. Cognitive science research on self-explanation identifies two mechanisms with the strongest empirical support: learners use the act of explaining to identify and fill knowledge gaps, and to construct or repair their underlying mental models. A multiple choice question cannot observe either mechanism. It only observes the output.

formative vs. summative assessment

Why the Format of the Example Matters More Than the Label

Calling something a formative assessment does not make it diagnostic. Assessment for learning requires that the act of assessing also builds understanding, not that it simply happens earlier in the term than a final exam. assessment for learning in practice

Axiom Flow is a direct implementation of assessment for learning: the act of teaching Sam, the platform's AI student, is itself the assessment, not preparation for one. Atlas, the assessment designer, generates a configurable set of misconceptions from the course material and maps each one to an exam question. Sam is initialized with those misconceptions and updates its understanding only based on the quality of the explanation it receives. What Sam still gets wrong after being taught is a direct trace of what the student never actually explained clearly.

This goes beyond the standard formative assessment platform model, which typically monitors whether a student picked the right answer on a quiz. As a teaching-based assessment, Axiom Flow measures whether the student can correct a specific, pre-mapped misconception through explanation, which is a much harder and more honest bar to clear.

A Better Standard for Choosing Formative Methods

When evaluating any formative method, including the examples above, the useful question is not "does this happen during learning." Nearly everything does. The useful question is whether the format requires the student to construct an explanation or merely recognize one.

Peer teaching, think-alouds, and structured concept explanation clear that bar. Polls, exit tickets built on short factual recall, and most quiz formats do not, even when they are well timed and frequently used. Recent systematic reviews of formative feedback in graduate programs point to the same pattern: the mechanism of the check matters as much as its frequency.

Instructors do not need to abandon quick checks. But if the goal is conceptual mastery assessment rather than a snapshot of recall, the format has to ask students to teach, explain, or defend, not just choose.