Formative Assessment in the Classroom: Evidence, Feedback and Next Steps

Use formative assessment to gather evidence during learning, interpret the answers and choose a next step the learner can use. A quiz becomes part of that process when its results lead to feedback, a changed explanation or new practice. This guide distinguishes formative and summative purposes and works through a fractions example, including a follow-up that tests whether the original reasoning has been overgeneralised.

What formative assessment actually is

The word “assessment” tends to summon images of red pens and end-of-term marks, but formative assessment is something quieter and more useful. Its purpose is not to certify learning after the fact; it is to inform what happens next. When a teacher poses a quick question, scans the room, reads a one-minute written response, or runs a brief quiz, the goal is diagnostic: what do my students grasp, and where are they stuck?

That information only counts as formative if it actually changes something. A number nobody acts on does not complete the formative loop; that alone does not make the task summative. The purpose and subsequent use of the evidence matter. The defining feature of formative assessment is the feedback loop: you gather evidence, you interpret it, and you adjust — reteaching a shaky concept, regrouping students, or moving on with confidence because the class is ready.

This distinction is grounded in a substantial research tradition rather than a label invented by Quiz software. Black and Wiliam’s foundational review, Assessment and Classroom Learning, examines how evidence is used to adapt teaching, and the OECD’s report on improving learning through formative assessment brings the approach into secondary-classroom practice. Neither implies that every quick Quiz works: the feedback loop and instructional response are the mechanism.

Formative vs summative: a difference of purpose, not format

It is tempting to sort assessments by their shape — a quiz is “formative,” an exam is “summative” — but the real distinction is purpose and timing.

• Summative assessment measures learning at the end of a unit, term, or course. It is a verdict: a grade, a certificate, a pass mark. Its job is to summarise.
• Formative assessment happens during learning. It is a compass: it points to what needs attention while there is still road ahead.

The same instrument can play either role. A ten-question Quiz can be summative if it is graded and final, or formative if its results feed straight back into teaching and the stakes are low. What matters is whether the results are used to adjust, not merely to record. Many strong teaching routines blend the two: frequent formative checks throughout a unit, and a single summative assessment at the end to confirm what has been learnt.

A worked loop: evidence, action and a new check

In this fictional class, you ask which is larger, 3/5 or 3/8 of the same whole, and request one reason. The correct answer is 3/5 because each fifth is larger than an eighth and both fractions contain three such parts. Before interpreting the distribution, check that the question states the same whole and that its answer key is correct.

If the follow-up remains confused, model another example and try again later. If explanations are secure, move to a task with a different numerator as well. These are teaching decisions based on evidence; they are not a claim that one quiz caused a measured improvement. The flexible grouping guide describes the separate task of organising temporary follow-up work.

Check whether the explanation generalises beyond one fraction rule

The worked pair above uses equal numerators. A learner might leave with the shortcut “the smaller denominator always means the larger fraction”, which fails when numerators differ. Plan the next question to distinguish the valid same-numerator reasoning from that overgeneralisation. The following cases use fractions of the same whole and are constructed for this example.

Suppose a fictional group succeeds on the first row but struggles on the second. Do not immediately conclude that the first lesson failed. Check whether equivalent fractions and common denominators have already been taught. If they have not, the second row introduces new content and needs modelling before independent assessment. If they have, read a few worked steps to distinguish a conversion error from a mistaken comparison rule.

Choose the response accordingly: revisit equivalent fractions, compare equal-sized visual models, or provide a new pair with a common denominator already shown. Record which support was supplied. A correct answer with that support documents a different performance from a correct answer without it, even if both receive encouragement.

Then remove the relevant support on a later, comparable pair and ask for a justification. This makes the formative loop observable: evidence from a particular task leads to a specific explanation and a new check. It avoids turning one wrong choice into a permanent learner label, or using a rising total as proof that one teaching move caused the change.

The role of feedback

Feedback is the step that connects the evidence to another attempt. Evidence of understanding is only valuable once it travels back to the learner in a form they can act on. Good formative feedback shares a few qualities:

• It is timely. Give feedback when the learner can use it in a further attempt; the appropriate timing depends on the task and the kind of feedback.
• It is specific. “Check your method for balancing the equation” is more useful than “Try harder.” Pinpointed feedback tells the student exactly what to fix.
• It is forward-looking. The best feedback describes the next step, not just the size of the gap. It answers the student’s real question: what do I do now?

Crucially, feedback flows both ways. Formative assessment tells students how they are doing, and it tells you how your teaching is landing. When half the class gets a question wrong, first inspect its wording, key and access conditions. Then examine explanations before deciding whether the concept needs another approach. When you write your checks carefully, you learn as much from the wrong answers as from the right ones, which is why it pays to write clear, diagnostic multiple-choice questions whose distractors map onto real misconceptions.

Differentiation: meeting students where they are

Responses can help you choose different follow-up tasks when learners need different support. Once you can see who has mastered a concept and who is still grappling with it, you can respond deliberately: extension tasks for those ready to push on, a small reteaching group for those who need another pass, a different worked example for the student who took a wrong turn three steps back.

Without ongoing evidence, differentiation collapses into assumption — teaching to an imagined “average” student who may not exist in your room. With it, you can group flexibly, vary the difficulty of follow-up tasks, and target your attention where it will do the most good. The point is not to label students but to keep adjusting the path so each one stays in productive difficulty: challenged, but not lost.

Choose a cadence that leaves time to act

Schedule a check when its answers can change a teaching decision. That may be at the start of a new topic, after an explanation or before a later revisit. More checks are not automatically better: each needs time for interpretation, feedback and another attempt. A short discussion or a worked example may be the more useful next step.

Keep the purpose clear and the stakes proportionate. Explain what you will do with the responses and offer a way to express uncertainty. This can lower pressure, but it does not guarantee honest answers or remove anxiety for every learner. Adjust the rhythm to the class and access needs rather than turning a suggested timetable into a rule.

How short, low-stakes quizzes serve formative goals

Of all the formative tools available, the short Quiz is among the most efficient, because it gives you structured evidence from every student at once — not just the three who always raise their hands. A well-placed five-question Quiz at the start of a lesson reveals what survived from last time; one at the end shows what stuck today.

The “low-stakes” part is what makes this work. Counting these checks lightly, or not at all, keeps the focus on understanding rather than scores. Explain that uncertainty is useful and show how you use the answers; a low-stakes label alone does not guarantee candid responses. A few principles help short quizzes pull their formative weight:

• Target one thing at a time. A check aimed at a single concept tells you precisely where the trouble is. A sprawling Quiz muddies the diagnosis.
• Vary the question type to the thinking you want. Recalling a definition, ordering steps in a process, matching terms, and reasoning through an open response each probe a different layer of understanding. Drawing on a range of formats — single and multiple choice, true/false, matching, ordering, numeric, fill-in-the-blank, open response — lets a check reach beyond surface recall.
• Read the pattern, not just the score. The value is in which questions tripped students up. Automatic marking and a simple breakdown of responses turn a pile of answers into a clear next move, instead of leaving you to tally by hand.

In Skolina, objective answers can be scored automatically after you check the key. Advanced question-level statistics require Starter, Pro or equivalent institutional access; open explanations still need teacher review. Students can participate through a shared link or QR code without creating an account. Try a five-question exit ticket on tomorrow’s lesson: choose one misconception, write two answers that expose it and decide in advance what you will reteach if many Students choose them. You can create a free teacher account to run that check, or use the companion guide on creating an online quiz for your classroom for the mechanics.

The current plans and limits show which Skolina features are included at each level. Check them before planning a repeated routine, especially if you need advanced statistics or anti-cheat signals rather than basic automatic correction.

A note of honesty on the AI features some tools, including Skolina, now offer: AI can draft questions from a topic or a set of notes, which saves time, but the generation runs on OpenAI’s models, meaning that text is processed outside the EU, and the output always needs a human read-through before it reaches students. AI is a drafting assistant for your formative checks, not an author you can trust unchecked.

Plan the next teaching step before collecting more answers

Before collecting another set of answers, write the decision they could change. In the fractions example, that might be whether to revisit equal-sized parts, teach equivalent fractions or ask for a more independent explanation. Leave time to inspect the key, consider access difficulties and let learners use the feedback. If no next step depends on the answers, a discussion or worked example may be a better use of that time.

If you would like to build a few of these checks for your own classroom, you can create a free teacher account and have your first low-stakes Quiz ready before your next lesson.