Formative Assessment: Evidence That Improves Learning

Make the learning target visible

Formative assessment is a cycle: clarify what learners should understand, elicit evidence while they are learning, interpret that evidence and use it to change the next move. The information can guide a teacher’s explanation, a learner’s revision or a peer discussion. A quiz is not formative merely because it is short or ungraded. If nobody uses its results while learning can still change, it has missed the defining purpose.

Begin with a specific target. “Understand fractions” is too broad for one check. “Explain why two fraction representations show the same quantity” names reasoning that can be observed. Show learners an example of a strong explanation and discuss why it works. A teacher can then ask a question that reveals whether learners recognize the same principle in a new case. The target and the evidence must fit; a multiple-choice answer alone may show selection without showing the explanation.

Decide what errors might mean. A learner who says two fourths is larger than one half because four is larger than two may be comparing denominators as whole numbers. Another may choose the correct answer by guessing. A third may understand the relation but write a confusing sentence. The teacher needs evidence that distinguishes these responses before choosing whether to revisit the model, ask a probing question or offer writing support.

Elicit evidence from more than volunteers

Choose a brief task that every learner can attempt. An individual diagram, written explanation, mini-whiteboard response or structured conversation can reveal different aspects of thinking. Asking “Does everyone get it?” produces little diagnostic information and invites a polite yes. A useful prompt requires learners to make a decision and give a reason. The teacher can scan answers for patterns, then invite explanations from a purposeful sample of learners.

Timing matters. A check before instruction reveals starting points; a check during guided practice shows whether to continue or slow down; an exit response informs the next lesson. Waiting until the end of a unit leaves fewer opportunities to act on a misconception. Use just enough evidence to make a decision. Constant interruptions for checking can crowd out the learning activity they are supposed to support.

The evidence also has limits. One correct answer does not prove durable understanding; one hesitant reply may reflect unfamiliar wording, stress or a misunderstanding of the question. Compare the response with earlier work, ask a follow-up and consider the task conditions. When a pattern appears across several learners, inspect whether the teacher’s example or the assessment prompt caused it. Avoid turning every error into a fixed judgment about ability.

Turn evidence into a response

Plan possible next steps before giving the prompt. If most learners compare the denominator alone, revisit equal-sized wholes with a visual model. If a smaller group is unsure, provide targeted practice while others explain a more demanding example. If learners can explain the idea accurately, introduce a transfer task. The response should address the reason for the mistake rather than simply present the correct answer again.

Feedback names what is working and what to do next. “Your diagram uses equal wholes; now explain why the shaded regions match” gives a learner a concrete action. “Try harder” does not. Allow time to use the feedback: redraw a model, revise a claim or solve a parallel problem. Feedback delivered after the task has ended may be ignored if the learner cannot apply it. Check a subsequent response to see whether the advice actually helped.

Peer discussion can expose thinking when the question is focused. Ask partners to compare reasons, identify the step where their explanations differ and revise a shared answer. Without clear criteria, discussion may merely exchange answers or let one student supply the work. The teacher still needs to monitor the quality of explanations and correct a misconception that spreads. Student self-assessment is useful when learners can compare their work with a known standard and choose a feasible next step.

Interpret results fairly

Make sure the task measures the intended idea rather than an avoidable obstacle. If the goal is mathematical reasoning, dense reading or unfamiliar cultural references may obscure what learners know. Offer appropriate access while retaining the same conceptual question. Record which supports were provided so the interpretation remains honest. A learner’s answer after extensive prompting conveys different information from an independent explanation.

Look beyond a single class average. A high percentage of correct choices may conceal fragile reasoning among some learners; an average decline may reflect a more demanding question. Inspect the kinds of answers, whose responses are missing and whether learners who need help receive it. Check whether a changed prompt is genuinely comparable to the earlier one before claiming improvement.

Privacy and trust shape what learners reveal. Public ranking of mistakes may discourage risk-taking. Use errors as evidence for teaching and discuss them without ridicule. An anonymous sample can help a class analyze an idea, while the teacher still follows up with individuals as needed. Explain how the information will be used so learners see the purpose of the check.

Close the loop and plan the next check

A written analysis can trace the full cycle through one concrete example: target, prompt, expected responses, interpretation, immediate adjustment and later evidence. Distinguish the observation from the inference. “Eight learners used the denominator as a whole number” describes responses; “they do not understand equivalent fractions” is a broader conclusion that requires more evidence. Name alternatives and the follow-up question that would test them.

Recheck after instruction in a fresh context. Learners might justify why three sixths equals one half without relying on the diagram used in the first lesson. If the explanation transfers, reduce support or move on. If it does not, revisit the underlying idea using a different representation. Do not claim success solely because the teacher delivered feedback; show what learners did with it.

Formative assessment is most useful when it is selective and consequential. A small, well-chosen question can change the direction of a lesson. Its value lies in the decision it improves and the opportunity it gives learners to revise their understanding before a final judgment is made.